Bump version: 0.38.0-beta.9 → 0.38.0-beta.10

2026-09-04 04:28:44 +00:00 · 2026-08-25 10:36:33 +00:00
100 changed files with 1105 additions and 9376 deletions
@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.38.0-beta.12"
+current_version = "0.38.0-beta.10"
 parse = """(?x)
    (?P<major>0|[1-9]\\d*)\\.
    (?P<minor>0|[1-9]\\d*)\\.
@@ -40,31 +40,40 @@ jobs:
          - target: aarch64-apple-darwin
            host: macos-latest
            features: fp16kernels
-            # Fat LTO was ~111 of this job's ~113 minutes.
-            lto: thin
-            codegen_units: 16
            pre_build: |-
              brew install protobuf
+              # Fat LTO (the workspace default in .cargo/config.toml) is
+              # single-threaded and is the peak-memory step of the build. On
+              # this runner it accounted for ~111 of the job's ~113 minutes,
+              # making it the critical path of the entire publish pipeline.
+              # ThinLTO parallelizes it across the runner's cores, for a few
+              # percent of runtime performance.
+              export CARGO_PROFILE_RELEASE_LTO=thin
+              export CARGO_PROFILE_RELEASE_CODEGEN_UNITS=16
          - target: x86_64-pc-windows-msvc
            host: windows-2025
            features: ","
-            # The lower peak also keeps this on the standard 4-core runner.
-            lto: thin
-            codegen_units: 16
            pre_build: |-
              choco install --no-progress protoc ninja nasm
              tail -n 1000 /c/ProgramData/chocolatey/logs/chocolatey.log
              # There is an issue where choco doesn't add nasm to the path
              export PATH="$PATH:/c/Program Files/NASM"
              nasm -v
+              # See the ThinLTO note on aarch64-apple-darwin above. Keeping
+              # peak memory down is also what lets this run on the standard
+              # 4-core runner: the 8-core larger runner was only needed to
+              # stop fat LTO from OOMing rustc-LLVM.
+              export CARGO_PROFILE_RELEASE_LTO=thin
+              export CARGO_PROFILE_RELEASE_CODEGEN_UNITS=16
          - target: aarch64-pc-windows-msvc
            host: windows-2025
            features: ","
-            lto: thin
-            codegen_units: 16
            pre_build: |-
                choco install --no-progress protoc
                rustup target add aarch64-pc-windows-msvc
+                # See the ThinLTO note on aarch64-apple-darwin above.
+                export CARGO_PROFILE_RELEASE_LTO=thin
+                export CARGO_PROFILE_RELEASE_CODEGEN_UNITS=16
          - target: x86_64-unknown-linux-gnu
            host: ubuntu-latest
            features: fp16kernels
@@ -94,14 +103,6 @@ jobs:
            # https://github.com/napi-rs/napi-rs/blob/main/debian-aarch64.Dockerfile
            docker: ghcr.io/napi-rs/napi-rs/nodejs-rust:lts-debian-aarch64
            features: "fp16kernels"
-            # Fat LTO OOM-killed rustc every nightly; even with lld it peaked
-            # at 31391 MiB of the runner's 32 GiB.
-            lto: thin
-            codegen_units: 16
-            # arm64 Linux links through GNU `ld` where x86_64 defaults to
-            # `rust-lld`, which is why only arm64 OOM'd. lld cut the largest
-            # linker process 7.0 -> 4.0 GiB (lancedb/sophon#7313).
-            linker: /tmp/aarch64-lld-clang
            pre_build: |-
              set -e &&
              apt-get update &&
@@ -111,30 +112,9 @@ jobs:
              # AT_HWCAP2 (added in Linux 3.17). Define it for aws-lc-sys.
              export CFLAGS="$CFLAGS -DAT_HWCAP2=26" &&
              rustup target add aarch64-unknown-linux-gnu
-              # Not `&&`-chained: in dash, errexit does not fire for a
-              # non-final command in an `&&` list, so failures were ignored.
-              #
-              # A wrapper rather than `-C link-arg` because the per-target
-              # rustflags variable does not reach every unit that links, while
-              # the linker variable does. `clang` because GCC silently ignores
-              # `-fuse-ld=lld` unless built with lld support. Two echoes
-              # because printf's newline escape gets rewritten to `;` between
-              # here and the container.
-              echo '#!/bin/sh' > /tmp/aarch64-lld-clang
-              echo 'exec clang --target=aarch64-unknown-linux-gnu --sysroot=/usr/aarch64-unknown-linux-gnu/aarch64-unknown-linux-gnu/sysroot --gcc-toolchain=/usr/aarch64-unknown-linux-gnu -fuse-ld=lld "$@"' >> /tmp/aarch64-lld-clang
-              chmod 0755 /tmp/aarch64-lld-clang
-              # Fail now, not at the cdylib link ~30 minutes later. Linking at
-              # all also proves lld resolved; clang errors out when it cannot.
-              echo 'int main(void){return 0;}' > /tmp/probe.c
-              /tmp/aarch64-lld-clang /tmp/probe.c -o /tmp/probe
-              readelf -h /tmp/probe | grep AArch64
          - target: aarch64-unknown-linux-musl
            host: ubuntu-2404-8x-x64
            features: ","
-            # Fat LTO took the whole runner down. lld cannot help: it died
-            # inside rustc's LLVM, before any linker was spawned.
-            lto: thin
-            codegen_units: 16
            pre_build: |-
              set -e &&
              sudo apt-get update &&
@@ -143,19 +123,6 @@ jobs:
              export EXTRA_ARGS="-x"
    name: build - ${{ matrix.settings.target }}
    runs-on: ${{ matrix.settings.host }}
-    # On the job, not exported from `pre_build`: `Swatinem/rust-cache` hashes
-    # `CARGO_*` into its cache key before any step runs, so a step-local export
-    # leaves the key unchanged while cargo still rebuilds cold. The ThinLTO
-    # legs had been doing that every run.
-    #
-    # Not `RUSTFLAGS`: setting it, even to "", discards every config-file
-    # rustflag, silently dropping .cargo/config.toml's `target-cpu` and
-    # `target-feature` from the published binaries.
-    env:
-      CARGO_PROFILE_RELEASE_LTO: ${{ matrix.settings.lto || 'fat' }}
-      CARGO_PROFILE_RELEASE_CODEGEN_UNITS: ${{ matrix.settings.codegen_units || '1' }}
-      # Empty elsewhere: a per-target variable is only read for that triple.
-      CARGO_TARGET_AARCH64_UNKNOWN_LINUX_GNU_LINKER: ${{ matrix.settings.linker }}
    defaults:
      run:
        working-directory: nodejs
@@ -202,15 +169,19 @@ jobs:
          # creating ref). The nightly cadence also keeps entries inside
          # GitHub's 7-day eviction window, which a tag-only trigger would not.
          save-if: ${{ github.ref == 'refs/heads/main' }}
-      # Docker builds can use rust-cache too: the workspace is bind-mounted, so
-      # `target/` lives on the host and rust-cache's prune keeps the entry
-      # small.
+      # Docker builds can use rust-cache too. `target/` already lives on the
+      # host because the whole workspace is bind-mounted into the container, and
+      # rust-cache's prune and save run host-side, so they can manage it -- which
+      # is what keeps the entry to dependency artifacts rather than a multi-GB
+      # copy of everything.
      #
      # Two differences from the native builds. The container's CARGO_HOME is
-      # bind-mounted from `.cargo-cache` rather than ~/.cargo, so that is cached
-      # explicitly. And the key uses the *host* rustc version, not the compiler
-      # that built these artifacts -- safe, since cargo fingerprints the real
-      # one; a base-image bump just costs one cold build.
+      # bind-mounted from `.cargo-cache` rather than the host's ~/.cargo, so that
+      # has to be cached explicitly. And the key is derived from the *host* rustc
+      # version, which is not the compiler that produced these artifacts; that is
+      # safe because cargo fingerprints the real compiler and rebuilds on a
+      # mismatch, it just means a base-image toolchain bump costs one cold build
+      # instead of invalidating the key.
      - name: Cache cargo (docker builds)
        uses: Swatinem/rust-cache@v2
        if: ${{ matrix.settings.docker }}
@@ -239,14 +210,9 @@ jobs:
          # cache step above saves. Previously the registry mounts pointed at
          # `.cargo/...`, a path nothing cached, so the container re-downloaded
          # the whole crate registry on every run.
-          #
-          # `docker run` inherits nothing; `-e NAME` carries the job's `env:` in.
          options: "--user 0:0 -v ${{ github.workspace }}/.cargo-cache/git/db:/usr/local/cargo/git/db \
                    -v ${{ github.workspace }}/.cargo-cache/registry/cache:/usr/local/cargo/registry/cache \
                    -v ${{ github.workspace }}/.cargo-cache/registry/index:/usr/local/cargo/registry/index \
-                    -e CARGO_PROFILE_RELEASE_LTO \
-                    -e CARGO_PROFILE_RELEASE_CODEGEN_UNITS \
-                    -e CARGO_TARGET_AARCH64_UNKNOWN_LINUX_GNU_LINKER \
                    -v ${{ github.workspace }}:/build -w /build/nodejs"
          run: |
            set -e
@@ -290,18 +256,6 @@ jobs:
        if: always()
        run: df -h
        shell: bash
-      - name: Report peak memory
-        if: always() && runner.os == 'Linux'
-        shell: bash
-        run: |
-          peak=$(find /sys/fs/cgroup -name memory.peak -readable \
-                 -exec cat {} + 2>/dev/null | sort -n | tail -1)
-          if [ -n "$peak" ]; then
-            echo "peak memory: $((peak / 1024 / 1024)) MiB"
-          else
-            echo "peak memory: unavailable (no readable cgroup v2 memory.peak)"
-          fi
-          free -g || true
      - name: Upload artifact
        uses: actions/upload-artifact@v7
        with:
@@ -1597,9 +1597,9 @@ checksum = "613afe47fcd5fac7ccf1db93babcb082c5994d996f20b8b159f2ad1658eb5724"

 [[package]]
 name = "chacha20"
-version = "0.10.2"
+version = "0.10.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "65c35e4b699c7e15ccbe7ee35c005e4fc0a278d22238a2857e6ce2dadeda1b06"
+checksum = "6f8d983286843e49675a4b7a2d174efe136dc93a18d69130dd18198a6c167601"
 dependencies = [
 "cfg-if 1.0.4",
 "cpufeatures 0.3.0",
@@ -3455,8 +3455,8 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"

 [[package]]
 name = "fsst"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "rand 0.9.5",
@@ -4815,8 +4815,8 @@ checksum = "e037a2e1d8d5fdbd49b16a4ea09d5d6401c1f29eca5ff29d03d3824dba16256a"

 [[package]]
 name = "lance"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arc-swap",
 "arrow",
@@ -4888,8 +4888,8 @@ dependencies = [

 [[package]]
 name = "lance-arrow"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4911,7 +4911,7 @@ dependencies = [
 [[package]]
 name = "lance-arrow-scalar"
 version = "58.0.0"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4925,7 +4925,7 @@ dependencies = [
 [[package]]
 name = "lance-arrow-stats"
 version = "58.0.0"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-schema",
@@ -4934,8 +4934,8 @@ dependencies = [

 [[package]]
 name = "lance-bitpacking"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrayref",
 "crunchy",
@@ -4945,8 +4945,8 @@ dependencies = [

 [[package]]
 name = "lance-core"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4983,8 +4983,8 @@ dependencies = [

 [[package]]
 name = "lance-datafusion"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -5013,8 +5013,8 @@ dependencies = [

 [[package]]
 name = "lance-datagen"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -5031,8 +5031,8 @@ dependencies = [

 [[package]]
 name = "lance-derive"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "proc-macro2",
 "quote",
@@ -5041,8 +5041,8 @@ dependencies = [

 [[package]]
 name = "lance-encoding"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-arith",
 "arrow-array",
@@ -5075,8 +5075,8 @@ dependencies = [

 [[package]]
 name = "lance-file"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-arith",
 "arrow-array",
@@ -5107,8 +5107,8 @@ dependencies = [

 [[package]]
 name = "lance-index"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arc-swap",
 "arrow",
@@ -5172,8 +5172,8 @@ dependencies = [

 [[package]]
 name = "lance-index-core"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-schema",
@@ -5195,8 +5195,8 @@ dependencies = [

 [[package]]
 name = "lance-io"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -5236,8 +5236,8 @@ dependencies = [

 [[package]]
 name = "lance-linalg"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-schema",
@@ -5251,8 +5251,8 @@ dependencies = [

 [[package]]
 name = "lance-namespace"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "async-trait",
@@ -5264,8 +5264,8 @@ dependencies = [

 [[package]]
 name = "lance-namespace-impls"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "arrow-ipc",
@@ -5318,8 +5318,8 @@ dependencies = [

 [[package]]
 name = "lance-select"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -5333,8 +5333,8 @@ dependencies = [

 [[package]]
 name = "lance-table"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -5374,8 +5374,8 @@ dependencies = [

 [[package]]
 name = "lance-testing"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "arrow-array",
 "arrow-schema",
@@ -5388,8 +5388,8 @@ dependencies = [

 [[package]]
 name = "lance-tokenizer"
-version = "12.0.0-beta.2"
-source = "git+https://github.com/lance-format/lance.git?tag=v12.0.0-beta.2#dafa4642658d996b3e31dde91e02f72db7860d7e"
+version = "11.0.0-beta.22"
+source = "git+https://github.com/lance-format/lance.git?tag=v11.0.0-beta.22#ea3cb4d799c468232735e9bcb43959487aca5c20"
 dependencies = [
 "frostem",
 "icu_segmenter",
@@ -5402,7 +5402,7 @@ dependencies = [

 [[package]]
 name = "lancedb"
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.9"
 dependencies = [
 "ahash",
 "anyhow",
@@ -5490,7 +5490,7 @@ dependencies = [

 [[package]]
 name = "lancedb-nodejs"
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.9"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -5515,7 +5515,7 @@ dependencies = [

 [[package]]
 name = "lancedb-python"
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.9"
 dependencies = [
 "arrow",
 "async-trait",
@@ -13,20 +13,20 @@ categories = ["database-implementations"]
 rust-version = "1.91.0"

 [workspace.dependencies]
-lance = { "version" = "=12.0.0-beta.2", default-features = false, "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=12.0.0-beta.2", default-features = false, "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=12.0.0-beta.2", default-features = false, "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=12.0.0-beta.2", "tag" = "v12.0.0-beta.2", "git" = "https://github.com/lance-format/lance.git" }
+lance = { "version" = "=11.0.0-beta.22", default-features = false, "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-core = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-datagen = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-file = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-io = { "version" = "=11.0.0-beta.22", default-features = false, "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-index = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-linalg = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace-impls = { "version" = "=11.0.0-beta.22", default-features = false, "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-table = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-testing = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-datafusion = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-encoding = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
+lance-arrow = { "version" = "=11.0.0-beta.22", "tag" = "v11.0.0-beta.22", "git" = "https://github.com/lance-format/lance.git" }
 lancedb = { path = "rust/lancedb", default-features = false }
 ahash = "0.8"
 # Note that this one does not include pyarrow
@@ -131,13 +131,18 @@ allow = [
    "BSD-3-Clause",
    "ISC",
    "Unicode-3.0",
+    "Unicode-DFS-2016",
    "Zlib",
    "CC0-1.0",
    "MPL-2.0",
    "BSL-1.0",
+    "OpenSSL",
    # 0BSD ("BSD Zero Clause") is effectively public domain — no attribution
    # required. Pulled in by `mock_instant`.
    "0BSD",
+    # bzip2-1.0.6 is the permissive upstream bzip2 license (BSD-like). Pulled
+    # in by `libbz2-rs-sys`, the pure-Rust bzip2 implementation.
+    "bzip2-1.0.6",
    # CDLA-Permissive-2.0 is a permissive data license used by `webpki-roots`
    # for the Mozilla CA root bundle. Data-only, distribution-compatible.
    "CDLA-Permissive-2.0",
@@ -145,7 +150,12 @@ allow = [
 confidence-threshold = 0.8
 # Per-crate license exceptions: allow a license for a specific crate only,
 # rather than globally via the `allow` list above.
-exceptions = []
+exceptions = [
+    # CDDL-1.0 (copyleft) is pulled in only as a dev/profiling dependency via
+    # `inferno` -> `pprof` -> `lance-testing`; it is a test dependency that we
+    # do not distribute, so scope the allowance to `inferno` alone.
+    { allow = ["CDDL-1.0"], crate = "inferno" },
+]
 # Crates whose license cannot be determined from Cargo metadata but whose
 # license we've manually confirmed from upstream. Keep this list minimal.
 [[licenses.clarify]]
@@ -14,7 +14,7 @@ Add the following dependency to your `pom.xml`:
 <dependency>
    <groupId>com.lancedb</groupId>
    <artifactId>lancedb-core</artifactId>
-    <version>0.38.0-beta.12</version>
+    <version>0.38.0-beta.10</version>
 </dependency>
 ```

@@ -1,518 +0,0 @@
-[**@lancedb/lancedb**](../README.md) • **Docs**
-
-***
-
-[@lancedb/lancedb](../globals.md) / AutoQuery
-
-# Class: AutoQuery
-
-A builder for automatic string searches.
-
-Automatic search determines whether to use full-text or vector search from
-the table revision selected for each execution. This builder exposes the
-common operations supported by both query families.
-
-## Extends
-
- `StandardQueryBase`&lt;`NativeQuery` \| `NativeVectorQuery`&gt;
-
-## Properties
-
-### inner
-
-```ts
-protected inner: Query | VectorQuery | Promise<Query | VectorQuery>;
-```
-
-#### Inherited from
-
-`StandardQueryBase.inner`
-
-## Methods
-
-### analyzePlan()
-
-```ts
-analyzePlan(distributedMetrics?): Promise<string>
-```
-
-Executes the query and returns the physical query plan annotated with runtime metrics.
-
-This is useful for debugging and performance analysis, as it shows how the query was executed
-and includes metrics such as elapsed time, rows processed, and I/O statistics.
-
-#### Parameters
-
-* **distributedMetrics?**: [`AnalyzePlanDistributedMetrics`](../type-aliases/AnalyzePlanDistributedMetrics.md)
-    How distributed worker metrics are displayed for remote query plans.
-    Defaults to `"aggregate"`.
-
-#### Returns
-
-`Promise`&lt;`string`&gt;
-
-A query execution plan with runtime metrics for each step.
-
-#### Example
-
-```ts
-import * as lancedb from "@lancedb/lancedb"
-
-const db = await lancedb.connect("./.lancedb");
-const table = await db.createTable("my_table", [
-  { vector: [1.1, 0.9], id: "1" },
-]);
-
-const plan = await table.query().nearestTo([0.5, 0.2]).analyzePlan();
-
-Example output (with runtime metrics inlined):
-AnalyzeExec verbose=true, metrics=[]
- ProjectionExec: expr=[id@3 as id, vector@0 as vector, _distance@2 as _distance], metrics=[output_rows=1, elapsed_compute=3.292µs]
-  Take: columns="vector, _rowid, _distance, (id)", metrics=[output_rows=1, elapsed_compute=66.001µs, batches_processed=1, bytes_read=8, iops=1, requests=1]
-   CoalesceBatchesExec: target_batch_size=1024, metrics=[output_rows=1, elapsed_compute=3.333µs]
-    GlobalLimitExec: skip=0, fetch=10, metrics=[output_rows=1, elapsed_compute=167ns]
-     FilterExec: _distance@2 IS NOT NULL, metrics=[output_rows=1, elapsed_compute=8.542µs]
-      SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], metrics=[output_rows=1, elapsed_compute=63.25µs, row_replacements=1]
-       KNNVectorDistance: metric=l2, metrics=[output_rows=1, elapsed_compute=114.333µs, output_batches=1]
-        LanceScan: uri=/path/to/data, projection=[vector], row_id=true, row_addr=false, ordered=false, metrics=[output_rows=1, elapsed_compute=103.626µs, bytes_read=549, iops=2, requests=2]
-```
-
-#### Inherited from
-
-`StandardQueryBase.analyzePlan`
-
-***
-
-### execute()
-
-```ts
-protected execute(options?): AsyncGenerator<RecordBatch<any>, void, unknown>
-```
-
-Execute the query and return the results as an
-
-#### Parameters
-
-* **options?**: `Partial`&lt;[`QueryExecutionOptions`](../interfaces/QueryExecutionOptions.md)&gt;
-
-#### Returns
-
-`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;
-
-#### See
-
- - AsyncIterator
-of
- - RecordBatch.
-
-By default, LanceDb will use many threads to calculate results and, when
-the result set is large, multiple batches will be processed at one time.
-This readahead is limited however and backpressure will be applied if this
-stream is consumed slowly (this constrains the maximum memory used by a
-single query)
-
-#### Inherited from
-
-`StandardQueryBase.execute`
-
-***
-
-### explainPlan()
-
-```ts
-explainPlan(verbose): Promise<string>
-```
-
-Generates an explanation of the query execution plan.
-
-#### Parameters
-
-* **verbose**: `boolean` = `false`
-    If true, provides a more detailed explanation. Defaults to false.
-
-#### Returns
-
-`Promise`&lt;`string`&gt;
-
-A Promise that resolves to a string containing the query execution plan explanation.
-
-#### Example
-
-```ts
-import * as lancedb from "@lancedb/lancedb"
-const db = await lancedb.connect("./.lancedb");
-const table = await db.createTable("my_table", [
-  { vector: [1.1, 0.9], id: "1" },
-]);
-const plan = await table.query().nearestTo([0.5, 0.2]).explainPlan();
-```
-
-#### Inherited from
-
-`StandardQueryBase.explainPlan`
-
-***
-
-### fastSearch()
-
-```ts
-fastSearch(): this
-```
-
-Skip searching un-indexed data. This can make search faster, but will miss
-any data that is not yet indexed.
-
-Use [Table#optimize](Table.md#optimize) to index all un-indexed data.
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.fastSearch`
-
-***
-
-### ~~filter()~~
-
-```ts
-filter(predicate): this
-```
-
-A filter statement to be applied to this query.
-
-#### Parameters
-
-* **predicate**: `string`
-
-#### Returns
-
-`this`
-
-#### See
-
-where
-
-#### Deprecated
-
-Use `where` instead
-
-#### Inherited from
-
-`StandardQueryBase.filter`
-
-***
-
-### fullTextSearch()
-
-```ts
-fullTextSearch(query, options?): this
-```
-
-#### Parameters
-
-* **query**: `string` \| [`FullTextQuery`](../interfaces/FullTextQuery.md)
-
-* **options?**: `Partial`&lt;[`FullTextSearchOptions`](../interfaces/FullTextSearchOptions.md)&gt;
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.fullTextSearch`
-
-***
-
-### limit()
-
-```ts
-limit(limit): this
-```
-
-Set the maximum number of results to return.
-
-By default, a plain search has no limit.  If this method is not
-called then every valid row from the table will be returned.
-
-#### Parameters
-
-* **limit**: `number`
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.limit`
-
-***
-
-### offset()
-
-```ts
-offset(offset): this
-```
-
-Set the number of rows to skip before returning results.
-
-This is useful for pagination.
-
-#### Parameters
-
-* **offset**: `number`
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.offset`
-
-***
-
-### orderBy()
-
-```ts
-orderBy(ordering): this
-```
-
-Sort the results by the specified column(s).
-
-#### Parameters
-
-* **ordering**: [`ColumnOrdering`](../interfaces/ColumnOrdering.md) \| [`ColumnOrdering`](../interfaces/ColumnOrdering.md)[]
-
-#### Returns
-
-`this`
-
-This query builder.
-
-#### Inherited from
-
-`StandardQueryBase.orderBy`
-
-***
-
-### outputSchema()
-
-```ts
-outputSchema(): Promise<Schema<any>>
-```
-
-Returns the schema of the output that will be returned by this query.
-
-This can be used to inspect the types and names of the columns that will be
-returned by the query before executing it.
-
-#### Returns
-
-`Promise`&lt;`Schema`&lt;`any`&gt;&gt;
-
-An Arrow Schema describing the output columns.
-
-#### Inherited from
-
-`StandardQueryBase.outputSchema`
-
-***
-
-### select()
-
-```ts
-select(columns): this
-```
-
-Return only the specified columns.
-
-By default a query will return all columns from the table.  However, this can have
-a very significant impact on latency.  LanceDb stores data in a columnar fashion.  This
-means we can finely tune our I/O to select exactly the columns we need.
-
-As a best practice you should always limit queries to the columns that you need.  If you
-pass in an array of column names then only those columns will be returned.
-
-You can also use this method to create new "dynamic" columns based on your existing columns.
-For example, you may not care about "a" or "b" but instead simply want "a + b".  This is often
-seen in the SELECT clause of an SQL query (e.g. `SELECT a+b FROM my_table`).
-
-To create dynamic columns you can pass in a Map<string, string>.  A column will be returned
-for each entry in the map.  The key provides the name of the column.  The value is
-an SQL string used to specify how the column is calculated.
-
-For example, an SQL query might state `SELECT a + b AS combined, c`.  The equivalent
-input to this method would be:
-
-#### Parameters
-
-* **columns**: `string` \| `string`[] \| `Record`&lt;`string`, `string`&gt; \| `Map`&lt;`string`, `string`&gt;
-
-#### Returns
-
-`this`
-
-#### Example
-
-```ts
-new Map([["combined", "a + b"], ["c", "c"]])
-
-Columns will always be returned in the order given, even if that order is different than
-the order used when adding the data.
-
-Note that you can pass in a `Record<string, string>` (e.g. an object literal). This method
-uses `Object.entries` which should preserve the insertion order of the object.  However,
-object insertion order is easy to get wrong and `Map` is more foolproof.
-```
-
-#### Inherited from
-
-`StandardQueryBase.select`
-
-***
-
-### toArray()
-
-```ts
-toArray(options?): Promise<any[]>
-```
-
-Collect the results as an array of objects.
-
-#### Parameters
-
-* **options?**: `Partial`&lt;[`QueryExecutionOptions`](../interfaces/QueryExecutionOptions.md)&gt;
-
-#### Returns
-
-`Promise`&lt;`any`[]&gt;
-
-#### Inherited from
-
-`StandardQueryBase.toArray`
-
-***
-
-### toArrow()
-
-```ts
-toArrow(options?): Promise<Table<any>>
-```
-
-Collect the results as an Arrow
-
-#### Parameters
-
-* **options?**: `Partial`&lt;[`QueryExecutionOptions`](../interfaces/QueryExecutionOptions.md)&gt;
-
-#### Returns
-
-`Promise`&lt;`Table`&lt;`any`&gt;&gt;
-
-#### See
-
-ArrowTable.
-
-#### Inherited from
-
-`StandardQueryBase.toArrow`
-
-***
-
-### useLsm()
-
-```ts
-useLsm(enable): this
-```
-
-Control MemWAL read routing for this query.
-
-By default (unset), when the table carries a MemWAL write spec (see
-[Table#setLsmWriteSpec](Table.md#setlsmwritespec)), reads are routed through the LSM scanner so
-they also return data written via the `mergeInsert` LSM path that has not yet
-been compacted into the base table (the active/frozen in-memory memtables and
-the flushed generations), deduplicated by primary key; a table without a spec
-reads the base table.
-
-#### Parameters
-
-* **enable**: `boolean`
-    `true` forces the LSM scanner and errors if the table has no
-    MemWAL write spec. `false` bypasses the MemWAL and reads the base table only,
-    even when a spec is present.
-    Note: the LSM scanner does not support every query shape (e.g. reranking,
-    hybrid search, `orderBy`). On a MemWAL table those shapes error unless
-    `useLsm(false)` is set, because a base-only read would silently exclude
-    un-compacted MemWAL data.
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.useLsm`
-
-***
-
-### where()
-
-```ts
-where(predicate): this
-```
-
-A filter statement to be applied to this query.
-
-The filter should be supplied as an SQL query string.  For example:
-
-#### Parameters
-
-* **predicate**: `string`
-
-#### Returns
-
-`this`
-
-#### Example
-
-```ts
-x > 10
-y > 0 AND y < 100
-x > 5 OR y = 'test'
-
-Filtering performance can often be improved by creating a scalar index
-on the filter column(s).
-
-Calling this multiple times combines the filters with a logical AND rather
-than replacing the previous filter.
-```
-
-#### Inherited from
-
-`StandardQueryBase.where`
-
-***
-
-### withRowId()
-
-```ts
-withRowId(): this
-```
-
-Whether to return the row id in the results.
-
-This column can be used to match results between different queries. For
-example, to match results from a full text search and a vector search in
-order to perform hybrid search.
-
-#### Returns
-
-`this`
-
-#### Inherited from
-
-`StandardQueryBase.withRowId`
@@ -942,7 +942,7 @@ Get the schema of the table.
 abstract search(
   query,
   queryType?,
-   ftsColumns?): Query | VectorQuery | AutoQuery
+   ftsColumns?): Query | VectorQuery
 ```

 Create a search query to find the nearest neighbors
@@ -964,7 +964,7 @@ of the given query

 #### Returns

-[`Query`](Query.md) \| [`VectorQuery`](VectorQuery.md) \| [`AutoQuery`](AutoQuery.md)
+[`Query`](Query.md) \| [`VectorQuery`](VectorQuery.md)

 ***

@@ -1292,18 +1292,6 @@ abstract updateFieldMetadata(updates): Promise<UpdateFieldMetadataResult>

 Update per-field (column) metadata.

-The following keys are treated specially, by convention, and should be
-used when appropriate:
-
- `lancedb:description`: for a human-readable description of a field.
- `lancedb:tag:<name>`: for a user-defined key-value tag, where the suffix
-  names the tag category; e.g. `lancedb:tag:model: "clip"`.
- `lancedb:logical-column`: for a column grouping; e.g. `feature_v1` and
-  `feature_v2` might be in the same logical column.
- `lancedb:status`: for status options (`production`, `candidate`,
-  `deprecated`, `archived`) to designate the current life cycle state of
-  this column.
-
 #### Parameters

 * **updates**: [`FieldMetadataUpdate`](../interfaces/FieldMetadataUpdate.md)[]
@@ -18,7 +18,6 @@

 ## Classes

- [AutoQuery](classes/AutoQuery.md)
 - [BooleanQuery](classes/BooleanQuery.md)
 - [BoostQuery](classes/BoostQuery.md)
 - [BranchContents](classes/BranchContents.md)
@@ -17,8 +17,7 @@ metadata: Record<string, null | string>;
 ```

 Metadata key/value pairs. Merged into the field's existing metadata by
-default; a value of `null` deletes that key. See
-[Table.updateFieldMetadata](../classes/Table.md#updatefieldmetadata) for the conventional `lancedb:*` keys.
+default; a value of `null` deletes that key.

 ***

@@ -10,12 +10,16 @@
 function getRegistry(): EmbeddingFunctionRegistry
 ```

-Get the global embedding function registry.
-
-LanceDB built-in providers are initialized when this public API is first
-used, so importing the root package does not change automatic search
-selection for tables without embedding metadata.
+Utility function to get the global instance of the registry

 ## Returns

 [`EmbeddingFunctionRegistry`](../classes/EmbeddingFunctionRegistry.md)
+
+`EmbeddingFunctionRegistry` The global instance of the registry
+
+## Example
+
+```ts
+const registry = getRegistry();
+const openai = registry.get("openai").create();
@@ -42,8 +42,6 @@ listing a storage directory.

 ::: lancedb.table.Table

-::: lancedb.table.CompactionOptions
-
 ::: lancedb.table.FragmentStatistics

 ::: lancedb.table.FragmentSummaryStats
@@ -161,8 +159,6 @@ and combined with [BooleanQuery][lancedb.query.BooleanQuery].

 ::: lancedb.query.FullTextOperator

-::: lancedb.query.DocumentGranularity
-
 ::: lancedb.query.Occur

 ## Embeddings
@@ -225,14 +221,10 @@ tokens = list(
 Blob columns store large binary values out of line so they can be read lazily
 instead of being materialized with the rest of the row.

-`lancedb.BlobType` is `lance.blob.BlobType` when pylance is installed. Without
-pylance, LanceDB uses a matching `lance.blob.v2` extension type so blob columns
-still work. Queries return descriptors. Call
-[`fetch_blob_files`][lancedb.table.Table.fetch_blob_files] for lazy reads or
-[`fetch_blobs`][lancedb.table.Table.fetch_blobs] for eager bytes.
-
 ::: lancedb.blob

+::: lancedb.BlobType
+
 ::: lancedb._blob.BlobFile
    options:
      show_root_full_path: false
@@ -8,7 +8,7 @@
    <parent>
      <groupId>com.lancedb</groupId>
      <artifactId>lancedb-parent</artifactId>
-      <version>0.38.0-beta.12</version>
+      <version>0.38.0-beta.10</version>
      <relativePath>../pom.xml</relativePath>
    </parent>

@@ -6,7 +6,7 @@

    <groupId>com.lancedb</groupId>
    <artifactId>lancedb-parent</artifactId>
-    <version>0.38.0-beta.12</version>
+    <version>0.38.0-beta.10</version>
    <packaging>pom</packaging>
    <name>${project.artifactId}</name>
    <description>LanceDB Java SDK Parent POM</description>
@@ -28,7 +28,7 @@
    <properties>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <arrow.version>15.0.0</arrow.version>
-        <lance-core.version>12.0.0-beta.2</lance-core.version>
+        <lance-core.version>11.0.0-beta.22</lance-core.version>
        <spotless.skip>false</spotless.skip>
        <spotless.version>2.30.0</spotless.version>
        <spotless.java.googlejavaformat.version>1.7</spotless.java.googlejavaformat.version>
@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.10"
 publish = false
 license.workspace = true
 description.workspace = true
@@ -1,16 +1,11 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
-import * as fs from "node:fs";
-import * as vm from "node:vm";
 import * as arrow15 from "apache-arrow-15";
 import * as arrow16 from "apache-arrow-16";
 import * as arrow17 from "apache-arrow-17";
 import * as arrow18 from "apache-arrow-18";

 import {
-  Field as CurrentField,
-  LargeBinary as CurrentLargeBinary,
-  Schema as CurrentSchema,
  Vector as CurrentVector,
  convertToTable,
  tableFromIPC as currentTableFromIPC,
@@ -41,59 +36,6 @@ function sampleRecords(): Array<Record<string, any>> {
    },
  ];
 }
-
-it("serializes an Arrow Table created in another JavaScript realm", async () => {
-  const context = vm.createContext({
-    TextDecoder,
-    TextEncoder,
-    console,
-    setTimeout,
-    clearTimeout,
-  });
-  vm.runInContext(
-    fs.readFileSync(
-      require.resolve("apache-arrow-15/Arrow.es2015.min"),
-      "utf8",
-    ),
-    context,
-  );
-  const foreignTable: unknown = vm.runInContext(
-    "Arrow.tableFromArrays({ id: new Int32Array([1, 2, 3]), text: ['foo', 'bar', 'baz'] })",
-    context,
-  );
-
-  const foreignMetadata = (
-    foreignTable as { schema: { metadata: Map<string, string> } }
-  ).schema.metadata;
-  expect(foreignMetadata).not.toBeInstanceOf(Map);
-
-  const buf = await fromDataToBuffer(
-    foreignTable as Parameters<typeof fromDataToBuffer>[0],
-  );
-  const actual = currentTableFromIPC(buf);
-
-  expect(actual.numRows).toBe(3);
-  expect(actual.getChild("id")?.toJSON()).toEqual([1, 2, 3]);
-  expect(actual.getChild("text")?.toJSON()).toEqual(["foo", "bar", "baz"]);
-});
-
-it("preserves field metadata from a provided schema", async function () {
-  const jsonMetadata = new Map([["ARROW:extension:name", "lance.json"]]);
-  const schema = new CurrentSchema([
-    new CurrentField("meta", new CurrentLargeBinary(), true, jsonMetadata),
-  ]);
-
-  const table = makeArrowTable(
-    [{ meta: Buffer.from(JSON.stringify({ source: "test" })) }],
-    { schema },
-  );
-
-  expect(table.schema.fields[0].metadata).toEqual(jsonMetadata);
-
-  const roundTripped = currentTableFromIPC(await fromTableToBuffer(table));
-  expect(roundTripped.schema.fields[0].metadata).toEqual(jsonMetadata);
-});
-
 describe.each([arrow15, arrow16, arrow17, arrow18])(
  "Arrow",
  (
@@ -187,58 +187,6 @@ describe("embedding functions", () => {
    const vector0 = JSON.parse(JSON.stringify(arr[0].vector));
    expect(vector0).toEqual([1, 2, 3]);
  });
-  it("should append multiple Python embeddings with the same alias", async () => {
-    @register("python-mock")
-    // biome-ignore lint/correctness/noUnusedVariables: the decorator registers this class
-    class MockEmbeddingFunction extends EmbeddingFunction<string> {
-      ndims() {
-        return 3;
-      }
-      embeddingDataType(): Float {
-        return new Float32();
-      }
-      async computeQueryEmbeddings(_data: string) {
-        return [1, 2, 3];
-      }
-      async computeSourceEmbeddings(data: string[]) {
-        return data.map((value) =>
-          value === "hello world" ? [1, 2, 3] : [4, 5, 6],
-        );
-      }
-    }
-
-    const metadata = new Map([
-      [
-        "embedding_functions",
-        '[{"source_column":"text1","vector_column":"vector1","name":"python-mock","model":{}},{"source_column":"text2","vector_column":"vector2","name":"python-mock","model":{}}]',
-      ],
-    ]);
-    const schema = new Schema(
-      [
-        new Field("text1", new Utf8(), true),
-        new Field("text2", new Utf8(), true),
-        new Field(
-          "vector1",
-          new FixedSizeList(3, new Field("item", new Float32(), true)),
-          true,
-        ),
-        new Field(
-          "vector2",
-          new FixedSizeList(3, new Field("item", new Float32(), true)),
-          true,
-        ),
-      ],
-      metadata,
-    );
-
-    const db = await connect(tmpDir.name);
-    const table = await db.createEmptyTable("test", schema);
-    await table.add([{ text1: "hello world", text2: "goodbye world" }]);
-
-    const rows = await table.query().toArray();
-    expect(JSON.parse(JSON.stringify(rows[0].vector1))).toEqual([1, 2, 3]);
-    expect(JSON.parse(JSON.stringify(rows[0].vector2))).toEqual([4, 5, 6]);
-  });

  it("should append generated vectors to a non-nullable schema", async () => {
    @register("non_nullable_schema_test")
@@ -1,95 +0,0 @@
-// SPDX-License-Identifier: Apache-2.0
-// SPDX-FileCopyrightText: Copyright The LanceDB Authors
-
-import { execFileSync } from "node:child_process";
-import { resolve } from "node:path";
-
-import type { OpenAIEmbeddingFunction } from "../lancedb/embedding/openai";
-import type { EmbeddingFunctionRegistry } from "../lancedb/embedding/registry";
-
-type EmbeddingModule = typeof import("../lancedb/embedding");
-type OpenAIModule = typeof import("../lancedb/embedding/openai");
-type RegistryModule = typeof import("../lancedb/embedding/registry");
-
-describe("embedding function registry", () => {
-  const registries: EmbeddingFunctionRegistry[] = [];
-
-  afterEach(() => {
-    for (const registry of registries) {
-      registry.reset();
-    }
-    registries.length = 0;
-  });
-
-  it("defers built-in providers until the public registry API is used", () => {
-    jest.isolateModules(() => {
-      const embedding = require("../lancedb/embedding") as EmbeddingModule;
-      const { getRegistry: getInternalRegistry } =
-        require("../lancedb/embedding/registry") as RegistryModule;
-      const registry = getInternalRegistry();
-      registries.push(registry);
-
-      expect(registry.length()).toBe(0);
-      expect(embedding.getRegistry()).toBe(registry);
-      expect(registry.get("openai")).toBeDefined();
-      expect(registry.get("huggingface")).toBeDefined();
-    });
-  });
-
-  it("preserves automatic FTS search in a fresh process", () => {
-    execFileSync(
-      process.execPath,
-      [resolve(__dirname, "fixtures", "auto_fts_search.cjs")],
-      { stdio: "pipe" },
-    );
-  });
-
-  it("shares registrations across duplicated provider module graphs", () => {
-    let registeringRegistry: EmbeddingFunctionRegistry | undefined;
-    let latestOpenAIConstructor: typeof OpenAIEmbeddingFunction | undefined;
-
-    jest.isolateModules(() => {
-      require("../lancedb/embedding/openai");
-      const { getRegistry } =
-        require("../lancedb/embedding/registry") as RegistryModule;
-      registeringRegistry = getRegistry();
-      registries.push(registeringRegistry);
-      expect(registeringRegistry.get("openai")).toBeDefined();
-    });
-
-    expect(() => {
-      jest.isolateModules(() => {
-        const { OpenAIEmbeddingFunction } =
-          require("../lancedb/embedding/openai") as OpenAIModule;
-        latestOpenAIConstructor = OpenAIEmbeddingFunction;
-        const { getRegistry } =
-          require("../lancedb/embedding/registry") as RegistryModule;
-        registries.push(getRegistry());
-      });
-    }).not.toThrow();
-
-    const previousApiKey = process.env.OPENAI_API_KEY;
-    process.env.OPENAI_API_KEY = "test";
-    try {
-      const latestOpenAI = registeringRegistry!
-        .get<OpenAIEmbeddingFunction>("openai")!
-        .create();
-      expect(latestOpenAI).toBeInstanceOf(latestOpenAIConstructor!);
-    } finally {
-      if (previousApiKey === undefined) {
-        delete process.env.OPENAI_API_KEY;
-      } else {
-        process.env.OPENAI_API_KEY = previousApiKey;
-      }
-    }
-
-    jest.isolateModules(() => {
-      const { getRegistry } =
-        require("../lancedb/embedding") as EmbeddingModule;
-      const publicRegistry = getRegistry();
-      registries.push(publicRegistry);
-      expect(publicRegistry).toBe(registeringRegistry);
-      expect(publicRegistry.get("openai")).toBeDefined();
-    });
-  });
-});
@@ -1,33 +0,0 @@
-// SPDX-License-Identifier: Apache-2.0
-// SPDX-FileCopyrightText: Copyright The LanceDB Authors
-
-const assert = require("node:assert/strict");
-const tmp = require("tmp");
-const { connect, embedding, Index } = require("../../dist");
-const { getRegistry } = require("../../dist/embedding/registry");
-
-async function main() {
-  assert.equal(typeof embedding.getRegistry, "function");
-  assert.equal(getRegistry().length(), 0);
-  assert.equal(embedding.getRegistry(), getRegistry());
-  assert.equal(getRegistry().length(), 2);
-
-  const dir = tmp.dirSync({ unsafeCleanup: true });
-  let db;
-  try {
-    db = await connect(dir.name);
-    const table = await db.createTable("docs", [{ text: "hello world" }]);
-    await table.createIndex("text", { config: Index.fts() });
-
-    const rows = await table.search("hello").toArray();
-    assert.equal(rows[0].text, "hello world");
-  } finally {
-    db?.close();
-    dir.removeCallback();
-  }
-}
-
-main().catch((error) => {
-  console.error(error);
-  process.exitCode = 1;
-});
@@ -11,13 +11,10 @@ import * as arrow17 from "apache-arrow-17";
 import * as arrow18 from "apache-arrow-18";

 import {
-  AutoQuery,
  Connection,
  MatchQuery,
  PhraseQuery,
-  Query,
  Table,
-  VectorQuery,
  connect,
  tokenize,
 } from "../lancedb";
@@ -685,56 +682,6 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
  },
 );

-// https://github.com/lancedb/lancedb/issues/1963
-it("should query documents with LangChain PDF metadata", async () => {
-  const tmpDir = tmp.dirSync({ unsafeCleanup: true });
-  try {
-    const db = await connect(tmpDir.name);
-    const documents = [
-      {
-        text: "first page",
-        vector: [1, 0],
-        source: "first.pdf",
-        loc: { pageNumber: 1, lines: { from: 1, to: 12 } },
-        pdf: {
-          version: "1.10.100",
-          info: {
-            format: "PDF 1.7",
-            producer: "pdf.js",
-            creator: "Writer",
-          },
-          totalPages: 2,
-        },
-      },
-      {
-        text: "second page",
-        vector: [0, 1],
-        source: "second.pdf",
-        loc: { pageNumber: 2, lines: { from: 13, to: 24 } },
-        pdf: {
-          version: "1.10.100",
-          info: {
-            format: "PDF 1.7",
-            producer: "pdf.js",
-            creator: "Writer",
-          },
-          totalPages: 2,
-        },
-      },
-    ];
-    const documentsTable = await db.createTable("documents", documents);
-
-    const results = await documentsTable.query().toArray();
-
-    expect(results).toHaveLength(2);
-    expect(results[0].source).toBe("first.pdf");
-    expect(results[0].pdf.info.producer).toBe("pdf.js");
-    expect(results[1].loc.pageNumber).toBe(2);
-  } finally {
-    tmpDir.removeCallback();
-  }
-});
-
 describe("merge insert", () => {
  let tmpDir: tmp.DirResult;
  let table: Table;
@@ -1830,194 +1777,6 @@ describe("Read consistency interval", () => {
  });
 });

-describe("automatic search schema consistency", () => {
-  let tmpDir: tmp.DirResult;
-
-  class SchemaRefreshEmbedding extends EmbeddingFunction<string> {
-    ndims() {
-      return 2;
-    }
-
-    embeddingDataType() {
-      return new Float32();
-    }
-
-    async computeSourceEmbeddings(data: string[]) {
-      return data.map((value) => [value.length, 1]);
-    }
-
-    async computeQueryEmbeddings(value: string) {
-      return [value.length, 1];
-    }
-  }
-
-  function embeddingSchema() {
-    const func = new SchemaRefreshEmbedding();
-    return LanceSchema({
-      text: func.sourceField(new Utf8()),
-      vector: func.vectorField(),
-    });
-  }
-
-  beforeEach(() => {
-    getRegistry().reset();
-    register("schema-refresh")(SchemaRefreshEmbedding);
-    tmpDir = tmp.dirSync({ unsafeCleanup: true });
-  });
-
-  afterEach(() => {
-    getRegistry().reset();
-    tmpDir.removeCallback();
-  });
-
-  it("uses the schema refreshed from another connection", async () => {
-    const first = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-    const second = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-
-    try {
-      const stale = await first.createTable("docs", [{ text: "before" }], {
-        schema: embeddingSchema(),
-      });
-      const replacement = await second.createTable(
-        "docs",
-        [{ text: "after hello" }],
-        { mode: "overwrite" },
-      );
-      await replacement.createIndex("text", { config: Index.fts() });
-
-      const search = stale.search("hello");
-      expect(search).toBeInstanceOf(AutoQuery);
-      expect(search).not.toBeInstanceOf(Query);
-      expect(search).not.toBeInstanceOf(VectorQuery);
-      expect("nprobes" in search).toBe(false);
-
-      const rows = await search.toArray();
-      expect(rows[0].text).toBe("after hello");
-      expect((await stale.schema()).metadata.has("embedding_functions")).toBe(
-        false,
-      );
-    } finally {
-      first.close();
-      second.close();
-    }
-  });
-
-  it("tracks embedding metadata across checkout and restore", async () => {
-    const first = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-    const second = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-
-    try {
-      await first.createTable("docs", [{ text: "before" }], {
-        schema: embeddingSchema(),
-      });
-      const table = await second.createTable(
-        "docs",
-        [{ text: "after hello" }],
-        { mode: "overwrite" },
-      );
-      await table.createIndex("text", { config: Index.fts() });
-
-      await table.checkout(1);
-      expect((await table.search("before").toArray())[0].text).toBe("before");
-
-      await table.checkoutLatest();
-      expect((await table.search("hello").toArray())[0].text).toBe(
-        "after hello",
-      );
-
-      await table.checkout(1);
-      await table.restore();
-      expect((await table.search("before").toArray())[0].text).toBe("before");
-    } finally {
-      first.close();
-      second.close();
-    }
-  });
-
-  it("pins automatic search while computing an embedding", async () => {
-    let markStarted!: () => void;
-    let releaseEmbedding!: () => void;
-    const started = new Promise<void>((resolve) => {
-      markStarted = resolve;
-    });
-    const released = new Promise<void>((resolve) => {
-      releaseEmbedding = resolve;
-    });
-
-    class BlockingEmbedding extends SchemaRefreshEmbedding {
-      async computeQueryEmbeddings(value: string) {
-        markStarted();
-        await released;
-        return [value.length, 1];
-      }
-    }
-
-    register("schema-refresh-blocking")(BlockingEmbedding);
-    const func = new BlockingEmbedding();
-    const schema = LanceSchema({
-      text: func.sourceField(new Utf8()),
-      vector: func.vectorField(),
-    });
-    const first = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-    const second = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-
-    try {
-      const table = await first.createTable(
-        "docs",
-        [{ text: "hello before" }],
-        { schema },
-      );
-      const pending = table.search("hello").toArray();
-      await started;
-
-      const replacement = await second.createTable(
-        "docs",
-        [{ text: "hello after" }],
-        { mode: "overwrite" },
-      );
-      await replacement.createIndex("text", { config: Index.fts() });
-      releaseEmbedding();
-
-      expect((await pending)[0].text).toBe("hello before");
-    } finally {
-      releaseEmbedding();
-      first.close();
-      second.close();
-    }
-  });
-
-  it("refreshes a reused automatic search for every execution", async () => {
-    const first = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-    const second = await connect(tmpDir.name, { readConsistencyInterval: 0 });
-
-    try {
-      const table = await first.createTable("docs", [
-        { text: "hello before", marker: "before" },
-      ]);
-      await table.createIndex("text", { config: Index.fts() });
-      const search = table.search("hello").select(["text"]);
-
-      const before = (await search.toArray())[0];
-      expect(before.text).toBe("hello before");
-      expect(before.marker).toBeUndefined();
-
-      const replacement = await second.createTable(
-        "docs",
-        [{ text: "hello after", marker: "after" }],
-        { mode: "overwrite" },
-      );
-      await replacement.createIndex("text", { config: Index.fts() });
-
-      const after = (await search.toArray())[0];
-      expect(after.text).toBe("hello after");
-      expect(after.marker).toBeUndefined();
-    } finally {
-      first.close();
-      second.close();
-    }
-  });
-});
-
 describe("schema evolution", function () {
  let tmpDir: tmp.DirResult;
  beforeEach(() => {
@@ -2585,24 +2344,7 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
      );
    });

-    test("full text search if only an unrelated embedding function is registered", async () => {
-      register("unused")(
-        class extends EmbeddingFunction<string> {
-          ndims() {
-            return 3;
-          }
-          embeddingDataType() {
-            return new Float32();
-          }
-          async computeQueryEmbeddings(_data: string) {
-            return [1, 2, 3];
-          }
-          async computeSourceEmbeddings(data: string[]) {
-            return data.map(() => [1, 2, 3]);
-          }
-        },
-      );
-
+    test("full text search if no embedding function provided", async () => {
      const db = await connect(tmpDir.name);
      const data = [
        { text: "hello world", vector: [0.1, 0.2, 0.3] },
@@ -2624,306 +2366,6 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
      expect(results2[0].text).toBe(data[1].text);
    });

-    test("auto search stays consistent with the active revision", async () => {
-      let initCalls = 0;
-      let queryCalls = 0;
-      let markStarted!: () => void;
-      const started = new Promise<void>((resolve) => {
-        markStarted = resolve;
-      });
-      let releaseEmbedding!: () => void;
-      const embeddingReleased = new Promise<void>((resolve) => {
-        releaseEmbedding = resolve;
-      });
-
-      @register("refresh-test")
-      class TestEmbedding extends EmbeddingFunction<string> {
-        async init() {
-          initCalls += 1;
-        }
-        ndims() {
-          return 1;
-        }
-        embeddingDataType() {
-          return new arrow.Float32();
-        }
-        async computeQueryEmbeddings(value: string) {
-          queryCalls += 1;
-          if (value === "blocked") {
-            markStarted();
-            await embeddingReleased;
-          }
-          return value === "greetings" ? [0.1] : [0.2];
-        }
-        async computeSourceEmbeddings(values: string[]) {
-          return values.map((value) =>
-            value === "hello world" ? [0.1] : [0.2],
-          );
-        }
-      }
-
-      const writer = await connect(tmpDir.name);
-      await writer.createTable("test", [{ text: "plain", vector: [0.0] }]);
-      const reader = await connect(tmpDir.name, {
-        readConsistencyInterval: 0,
-      });
-      const tracked = await reader.openTable("test");
-      type SnapshotCountingNative = {
-        querySnapshot: () => Promise<unknown>;
-      };
-      const native = (tracked as unknown as { inner: SnapshotCountingNative })
-        .inner;
-      const querySnapshot = native.querySnapshot.bind(native);
-      let snapshotCalls = 0;
-      native.querySnapshot = async () => {
-        snapshotCalls += 1;
-        return await querySnapshot();
-      };
-      const autoQuery = tracked.search("greetings").select(["text"]).limit(1);
-
-      const func = new TestEmbedding();
-      const schema = LanceSchema({
-        text: func.sourceField(new arrow.Utf8()),
-        vector: func.vectorField(),
-      });
-      const data = [{ text: "hello world" }, { text: "goodbye world" }];
-      await writer.createTable("test", data, { mode: "overwrite", schema });
-      const baselineInitCalls = initCalls;
-
-      expect(
-        (await tracked.schema()).metadata.get("embedding_functions"),
-      ).toBeDefined();
-      const results = await autoQuery.toArray();
-      expect(results[0].text).toBe(data[0].text);
-      expect(initCalls).toBe(baselineInitCalls + 1);
-      expect(queryCalls).toBe(1);
-      expect(snapshotCalls).toBe(1);
-
-      const repeatedResults = await autoQuery.toArray();
-      expect(repeatedResults[0].text).toBe(data[0].text);
-      expect(initCalls).toBe(baselineInitCalls + 1);
-      expect(queryCalls).toBe(1);
-      expect(snapshotCalls).toBe(2);
-
-      const pending = tracked
-        .search("blocked")
-        .select(["text"])
-        .limit(1)
-        .toArray();
-      await started;
-
-      const ftsData = [
-        { text: "greetings from full text", vector: [0.0] },
-        { text: "blocked from full text", vector: [0.0] },
-      ];
-      const ftsTable = await writer.createTable("test", ftsData, {
-        mode: "overwrite",
-      });
-      await ftsTable.createIndex("text", { config: Index.fts() });
-      releaseEmbedding();
-
-      const pendingResults = await pending;
-      expect(pendingResults[0].text).toBe(data[1].text);
-
-      expect(
-        (await tracked.schema()).metadata.get("embedding_functions"),
-      ).toBeUndefined();
-      const ftsResults = await autoQuery.toArray();
-      expect(ftsResults[0].text).toBe(ftsData[0].text);
-    });
-
-    test("auto search keeps newer preparation during a revision race", async () => {
-      let aCalls = 0;
-      let bCalls = 0;
-      let markAStarted!: () => void;
-      const aStarted = new Promise<void>((resolve) => {
-        markAStarted = resolve;
-      });
-      let releaseA!: () => void;
-      const aReleased = new Promise<void>((resolve) => {
-        releaseA = resolve;
-      });
-      let markBStarted!: () => void;
-      const bStarted = new Promise<void>((resolve) => {
-        markBStarted = resolve;
-      });
-      let releaseB!: () => void;
-      const bReleased = new Promise<void>((resolve) => {
-        releaseB = resolve;
-      });
-
-      @register("race-a")
-      class EmbeddingA extends EmbeddingFunction<string> {
-        ndims() {
-          return 1;
-        }
-        embeddingDataType() {
-          return new arrow.Float32();
-        }
-        async computeQueryEmbeddings() {
-          aCalls += 1;
-          markAStarted();
-          await aReleased;
-          return [0.1];
-        }
-        async computeSourceEmbeddings(values: string[]) {
-          return values.map(() => [0.1]);
-        }
-      }
-
-      @register("race-b")
-      class EmbeddingB extends EmbeddingFunction<string> {
-        ndims() {
-          return 1;
-        }
-        embeddingDataType() {
-          return new arrow.Float32();
-        }
-        async computeQueryEmbeddings() {
-          bCalls += 1;
-          markBStarted();
-          await bReleased;
-          return [0.2];
-        }
-        async computeSourceEmbeddings(values: string[]) {
-          return values.map(() => [0.2]);
-        }
-      }
-
-      const writer = await connect(tmpDir.name);
-      const embeddingA = new EmbeddingA();
-      const schemaA = LanceSchema({
-        text: embeddingA.sourceField(new arrow.Utf8()),
-        vector: embeddingA.vectorField(),
-      });
-      await writer.createTable("race", [{ text: "revision a" }], {
-        schema: schemaA,
-      });
-      const reader = await connect(tmpDir.name, {
-        readConsistencyInterval: 0,
-      });
-      const tracked = await reader.openTable("race");
-      const query = tracked.search("query");
-
-      const first = query.toArray();
-      await aStarted;
-
-      const embeddingB = new EmbeddingB();
-      const schemaB = LanceSchema({
-        text: embeddingB.sourceField(new arrow.Utf8()),
-        vector: embeddingB.vectorField(),
-      });
-      await writer.createTable("race", [{ text: "revision b" }], {
-        mode: "overwrite",
-        schema: schemaB,
-      });
-      const second = query.toArray();
-      await bStarted;
-
-      releaseA();
-      releaseB();
-      await Promise.all([first, second]);
-      expect(aCalls).toBe(1);
-      expect(bCalls).toBe(1);
-    });
-
-    test("stale FTS routing keeps newer vector preparation", async () => {
-      let vectorCalls = 0;
-      let markVectorStarted!: () => void;
-      const vectorStarted = new Promise<void>((resolve) => {
-        markVectorStarted = resolve;
-      });
-      let releaseVector!: () => void;
-      const vectorReleased = new Promise<void>((resolve) => {
-        releaseVector = resolve;
-      });
-
-      @register("stale-fts-race")
-      class RaceEmbedding extends EmbeddingFunction<string> {
-        ndims() {
-          return 1;
-        }
-        embeddingDataType() {
-          return new arrow.Float32();
-        }
-        async computeQueryEmbeddings() {
-          vectorCalls += 1;
-          markVectorStarted();
-          await vectorReleased;
-          return [0.1];
-        }
-        async computeSourceEmbeddings(values: string[]) {
-          return values.map(() => [0.1]);
-        }
-      }
-
-      const writer = await connect(tmpDir.name);
-      const ftsTable = await writer.createTable("stale_fts", [
-        { text: "hello", vector: [0.0] },
-      ]);
-      await ftsTable.createIndex("text", { config: Index.fts() });
-
-      const reader = await connect(tmpDir.name, {
-        readConsistencyInterval: 0,
-      });
-      const tracked = await reader.openTable("stale_fts");
-      type Snapshot = {
-        schema: () => Promise<Buffer>;
-      };
-      type NativeWithSnapshot = {
-        querySnapshot: () => Promise<Snapshot>;
-      };
-      const native = (tracked as unknown as { inner: NativeWithSnapshot })
-        .inner;
-      const querySnapshot = native.querySnapshot.bind(native);
-      let snapshotCalls = 0;
-      let markStaleSchemaStarted!: () => void;
-      const staleSchemaStarted = new Promise<void>((resolve) => {
-        markStaleSchemaStarted = resolve;
-      });
-      let releaseStaleSchema!: () => void;
-      const staleSchemaReleased = new Promise<void>((resolve) => {
-        releaseStaleSchema = resolve;
-      });
-      native.querySnapshot = async () => {
-        const snapshot = await querySnapshot();
-        snapshotCalls += 1;
-        if (snapshotCalls === 1) {
-          const schema = snapshot.schema.bind(snapshot);
-          snapshot.schema = async () => {
-            markStaleSchemaStarted();
-            await staleSchemaReleased;
-            return await schema();
-          };
-        }
-        return snapshot;
-      };
-
-      const query = tracked.search("hello");
-      const staleFtsExecution = query.toArray();
-      await staleSchemaStarted;
-
-      const embedding = new RaceEmbedding();
-      const vectorSchema = LanceSchema({
-        text: embedding.sourceField(new arrow.Utf8()),
-        vector: embedding.vectorField(),
-      });
-      await writer.createTable("stale_fts", [{ text: "hello" }], {
-        mode: "overwrite",
-        schema: vectorSchema,
-      });
-
-      const vectorExecution = query.toArray();
-      await vectorStarted;
-      releaseStaleSchema();
-      await staleFtsExecution;
-      releaseVector();
-      await vectorExecution;
-
-      await query.toArray();
-      expect(vectorCalls).toBe(1);
-    });
-
    test("tokenizes FTS queries by column or index name", async () => {
      const db = await connect(tmpDir.name);
      const data = [
@@ -3474,30 +2916,6 @@ describe("column name options", () => {
    expect(results[1].query_index).toBe(1);
  });

-  test("observes promised additional vectors while the query is pending", async () => {
-    const initialVector = new Promise<number[]>(() => undefined);
-    const query = table.query().nearestTo(initialVector);
-    const unhandled: unknown[] = [];
-    const onUnhandled = (reason: unknown) => unhandled.push(reason);
-    process.on("unhandledRejection", onUnhandled);
-
-    try {
-      query.addQueryVector(Promise.reject(new Error("extra vector failed")));
-      await new Promise<void>((resolve) => setImmediate(resolve));
-      expect(unhandled).toEqual([]);
-
-      const rejectedQuery = table
-        .query()
-        .nearestTo([0.1, 0.2])
-        .addQueryVector(Promise.reject(new Error("consumed vector failed")));
-      await expect(rejectedQuery.toArray()).rejects.toThrow(
-        "consumed vector failed",
-      );
-    } finally {
-      process.off("unhandledRejection", onUnhandled);
-    }
-  });
-
  test("index and search multivectors", async () => {
    const db = await connect(tmpDir.name);
    const data = [];
@@ -3561,27 +2979,6 @@ describe("when creating an empty table", () => {
    expect((actualSchema.fields[1].type as Float64).precision).toBe(2);
  });

-  it("can add and query JSON data", async () => {
-    const schema = new Schema([
-      new Field("id", new Int32(), true),
-      new Field(
-        "meta",
-        new Utf8(),
-        true,
-        new Map([["ARROW:extension:name", "arrow.json"]]),
-      ),
-    ]);
-    const table = await con.createEmptyTable("json", schema);
-    const meta = JSON.stringify({ x: 1 });
-
-    await table.add([{ id: 1, meta }]);
-
-    const rows = await table.query().toArray();
-    expect(rows).toHaveLength(1);
-    expect(rows[0].id).toBe(1);
-    expect(rows[0].meta).toBe(meta);
-  });
-
  it("can create an empty table from schema that specifies field types by name", async () => {
    const schemaLike = {
      fields: [
@@ -170,7 +170,7 @@ test("basic table examples", async () => {
    // --8<-- [end:create_index]

    // --8<-- [start:delete_rows]
-    await tbl.delete("item = 'fizz'");
+    await tbl.delete('item = "fizz"');
    // --8<-- [end:delete_rows]

    // --8<-- [start:drop_table]
@@ -72,7 +72,8 @@ export type FieldLike =
    };

 export type DataLike =
-  | import("apache-arrow").Data
+  // biome-ignore lint/suspicious/noExplicitAny: <explanation>
+  | import("apache-arrow").Data<Struct<any>>
  | {
      // biome-ignore lint/suspicious/noExplicitAny: <explanation>
      type: any;
@@ -81,7 +82,6 @@ export type DataLike =
      stride: number;
      nullable: boolean;
      children: DataLike[];
-      dictionary?: { data: readonly DataLike[] };
      get nullCount(): number;
      // biome-ignore lint/suspicious/noExplicitAny: <explanation>
      values: Buffers<any>[BufferType.DATA];
@@ -4,15 +4,7 @@
 import { Field, Schema } from "../arrow";
 import { sanitizeType } from "../sanitize";
 import { EmbeddingFunction } from "./embedding_function";
-import {
-  EmbeddingFunctionConfig,
-  EmbeddingFunctionRegistry,
-  getRegistry as getGlobalRegistry,
-  registerBuiltIn,
-} from "./registry";
-
-type OpenAIModule = typeof import("./openai");
-type TransformersModule = typeof import("./transformers");
+import { EmbeddingFunctionConfig, getRegistry } from "./registry";

 export {
  FieldOptions,
@@ -22,39 +14,7 @@ export {
  EmbeddingFunctionConstructor,
 } from "./embedding_function";

-export {
-  EmbeddingFunctionRegistry,
-  parseEmbeddingMetadata,
-  register,
-} from "./registry";
-export type {
-  CreateReturnType,
-  EmbeddingFunctionConfig,
-  EmbeddingFunctionCreate,
-  EmbeddingMetadataEntry,
-  ResolvedEmbeddingFunctionConfig,
-} from "./registry";
-
-function initializeBuiltInProviders() {
-  const { OpenAIEmbeddingFunction } = require("./openai") as OpenAIModule;
-  const { TransformersEmbeddingFunction } =
-    require("./transformers") as TransformersModule;
-
-  registerBuiltIn("openai", OpenAIEmbeddingFunction);
-  registerBuiltIn("huggingface", TransformersEmbeddingFunction);
-}
-
-/**
- * Get the global embedding function registry.
- *
- * LanceDB built-in providers are initialized when this public API is first
- * used, so importing the root package does not change automatic search
- * selection for tables without embedding metadata.
- */
-export function getRegistry(): EmbeddingFunctionRegistry {
-  initializeBuiltInProviders();
-  return getGlobalRegistry();
-}
+export * from "./registry";

 /**
 * Create a schema with embedding functions.
@@ -5,13 +5,14 @@ import type OpenAI from "openai";
 import type { EmbeddingCreateParams } from "openai/resources/index";
 import { Float, Float32 } from "../arrow";
 import { EmbeddingFunction } from "./embedding_function";
-import { registerBuiltIn } from "./registry";
+import { register } from "./registry";

 export type OpenAIOptions = {
  apiKey: string;
  model: EmbeddingCreateParams["model"];
 };

+@register("openai")
 export class OpenAIEmbeddingFunction extends EmbeddingFunction<
  string,
  Partial<OpenAIOptions>
@@ -99,5 +100,3 @@ export class OpenAIEmbeddingFunction extends EmbeddingFunction<
    return response.data[0].embedding;
  }
 }
-
-registerBuiltIn("openai", OpenAIEmbeddingFunction);
@@ -7,10 +7,6 @@ import {
 } from "./embedding_function";
 import "reflect-metadata";

-const builtInFunctionsKey = Symbol.for(
-  "@lancedb/lancedb::embedding-built-in-functions::v1",
-);
-
 export type CreateReturnType<T> = T extends { init: () => Promise<void> }
  ? Promise<T>
  : T;
@@ -63,15 +59,6 @@ export class EmbeddingFunctionRegistry {
    };
  }

-  /** @ignore */
-  setBuiltIn<
-    T extends EmbeddingFunctionConstructor = EmbeddingFunctionConstructor,
-  >(name: string, ctor: T): T {
-    this.#functions.set(name, ctor);
-    Reflect.defineMetadata("lancedb::embedding::name", name, ctor);
-    return ctor;
-  }
-
  get<T extends EmbeddingFunction<unknown>>(
    name: string,
  ): EmbeddingFunctionCreate<T> | undefined;
@@ -109,7 +96,6 @@ export class EmbeddingFunctionRegistry {
   */
  reset(this: EmbeddingFunctionRegistry) {
    this.#functions.clear();
-    getBuiltInFunctions(this).clear();
  }

  /**
@@ -197,56 +183,12 @@ export class EmbeddingFunctionRegistry {
  }
 }

-function getBuiltInFunctions(registry: EmbeddingFunctionRegistry): Set<string> {
-  const registryWithBuiltIns = registry as EmbeddingFunctionRegistry & {
-    [key: symbol]: Set<string> | undefined;
-  };
-  let builtInFunctions = registryWithBuiltIns[builtInFunctionsKey];
-  if (builtInFunctions === undefined) {
-    builtInFunctions = new Set<string>();
-    registryWithBuiltIns[builtInFunctionsKey] = builtInFunctions;
-  }
-  return builtInFunctions;
-}
-
-// Server bundlers can load the side-effect embedding entry points and the public
-// embedding API from separate module graphs. Keep their registry shared.
-const registryKey = Symbol.for(
-  "@lancedb/lancedb::embedding-function-registry::v1",
-);
-const registryGlobal = globalThis as typeof globalThis & {
-  [key: symbol]: EmbeddingFunctionRegistry | undefined;
-};
-
-function getGlobalRegistry(): EmbeddingFunctionRegistry {
-  const existingRegistry = registryGlobal[registryKey];
-  if (existingRegistry !== undefined) {
-    return existingRegistry;
-  }
-  const registry = new EmbeddingFunctionRegistry();
-  registryGlobal[registryKey] = registry;
-  return registry;
-}
-
-const _REGISTRY = getGlobalRegistry();
+const _REGISTRY = new EmbeddingFunctionRegistry();

 export function register(name?: string) {
  return _REGISTRY.register(name);
 }

-/** @ignore */
-export function registerBuiltIn<
-  T extends EmbeddingFunctionConstructor = EmbeddingFunctionConstructor,
->(name: string, ctor: T): T {
-  const builtInFunctions = getBuiltInFunctions(_REGISTRY);
-  if (builtInFunctions.has(name)) {
-    return _REGISTRY.setBuiltIn(name, ctor);
-  }
-  _REGISTRY.register(name)(ctor);
-  builtInFunctions.add(name);
-  return ctor;
-}
-
 /**
 * Utility function to get the global instance of the registry
 * @returns `EmbeddingFunctionRegistry` The global instance of the registry
@@ -3,7 +3,7 @@

 import { Float, Float32 } from "../arrow";
 import { EmbeddingFunction } from "./embedding_function";
-import { registerBuiltIn } from "./registry";
+import { register } from "./registry";

 export type XenovaTransformerOptions = {
  /** The wasm compatible model to use */
@@ -31,6 +31,7 @@ export type XenovaTransformerOptions = {
  };
 };

+@register("huggingface")
 export class TransformersEmbeddingFunction extends EmbeddingFunction<
  string,
  Partial<XenovaTransformerOptions>
@@ -157,8 +158,6 @@ export class TransformersEmbeddingFunction extends EmbeddingFunction<
  }
 }

-registerBuiltIn("huggingface", TransformersEmbeddingFunction);
-
 const tensorDiv = (
  src: import("@huggingface/transformers").Tensor,
  divBy: number,
@@ -103,7 +103,6 @@ export {
 } from "./native.js";

 export {
-  AutoQuery,
  ExecutableQuery,
  Query,
  QueryBase,
@@ -100,29 +100,6 @@ export interface FullTextSearchOptions {
  columns?: string | string[];
 }

-function nearestToNative(
-  inner: NativeQuery,
-  vector: Awaited<IntoVector>,
-): NativeVectorQuery {
-  const raw = Array.isArray(vector) ? null : extractVectorBuffer(vector);
-  if (raw) {
-    return inner.nearestToRaw(raw.data, raw.dtype);
-  }
-  return inner.nearestTo(Float32Array.from(vector as number[]));
-}
-
-function addQueryVectorToNative(
-  inner: NativeVectorQuery,
-  vector: Awaited<IntoVector>,
-) {
-  const raw = Array.isArray(vector) ? null : extractVectorBuffer(vector);
-  if (raw) {
-    inner.addQueryVectorRaw(raw.data, raw.dtype);
-  } else {
-    inner.addQueryVector(Float32Array.from(vector as number[]));
-  }
-}
-
 /** Common methods supported by all query types
 *
 * @see {@link Query}
@@ -134,15 +111,13 @@ export class QueryBase<
  NativeQueryType extends NativeQuery | NativeVectorQuery | NativeTakeQuery,
 > implements AsyncIterable<RecordBatch>
 {
-  protected inner!: NativeQueryType | Promise<NativeQueryType>;
-
  /**
   * @hidden
   */
-  protected constructor(inner?: NativeQueryType | Promise<NativeQueryType>) {
-    if (inner !== undefined) {
-      this.inner = inner;
-    }
+  protected constructor(
+    protected inner: NativeQueryType | Promise<NativeQueryType>,
+  ) {
+    // intentionally empty
  }

  // call a function on the inner (either a promise or the actual object)
@@ -160,15 +135,6 @@ export class QueryBase<
    }
  }

-  /**
-   * Return the native query used by the next terminal operation.
-   *
-   * @hidden
-   */
-  protected async getInner(): Promise<NativeQueryType> {
-    return this.inner;
-  }
-
  /**
   * Return only the specified columns.
   *
@@ -241,11 +207,16 @@ export class QueryBase<
  /**
   * @hidden
   */
-  protected async nativeExecute(
+  protected nativeExecute(
    options?: Partial<QueryExecutionOptions>,
  ): Promise<NativeBatchIterator> {
-    const inner = await this.getInner();
-    return inner.execute(options?.maxBatchLength, options?.timeoutMs);
+    if (this.inner instanceof Promise) {
+      return this.inner.then((inner) =>
+        inner.execute(options?.maxBatchLength, options?.timeoutMs),
+      );
+    } else {
+      return this.inner.execute(options?.maxBatchLength, options?.timeoutMs);
+    }
  }

  /**
@@ -274,7 +245,12 @@ export class QueryBase<
  /** Collect the results as an Arrow @see {@link ArrowTable}. */
  async toArrow(options?: Partial<QueryExecutionOptions>): Promise<ArrowTable> {
    const batches = [];
-    const inner = await this.getInner();
+    let inner;
+    if (this.inner instanceof Promise) {
+      inner = await this.inner;
+    } else {
+      inner = this.inner;
+    }
    for await (const batch of new RecordBatchIterable(inner, options)) {
      batches.push(batch);
    }
@@ -303,8 +279,11 @@ export class QueryBase<
   * @returns A Promise that resolves to a string containing the query execution plan explanation.
   */
  async explainPlan(verbose = false): Promise<string> {
-    const inner = await this.getInner();
-    return inner.explainPlan(verbose);
+    if (this.inner instanceof Promise) {
+      return this.inner.then((inner) => inner.explainPlan(verbose));
+    } else {
+      return this.inner.explainPlan(verbose);
+    }
  }

  /**
@@ -342,8 +321,13 @@ export class QueryBase<
    distributedMetrics?: AnalyzePlanDistributedMetrics,
  ): Promise<string> {
    const distributedMetricsMode = distributedMetrics ?? "aggregate";
-    const inner = await this.getInner();
-    return inner.analyzePlan(distributedMetricsMode);
+    if (this.inner instanceof Promise) {
+      return this.inner.then((inner) =>
+        inner.analyzePlan(distributedMetricsMode),
+      );
+    } else {
+      return this.inner.analyzePlan(distributedMetricsMode);
+    }
  }

  /**
@@ -355,8 +339,12 @@ export class QueryBase<
   * @returns An Arrow Schema describing the output columns.
   */
  async outputSchema(): Promise<import("./arrow").Schema> {
-    const inner = await this.getInner();
-    const schemaBuffer = await inner.outputSchema();
+    let schemaBuffer: Buffer;
+    if (this.inner instanceof Promise) {
+      schemaBuffer = await this.inner.then((inner) => inner.outputSchema());
+    } else {
+      schemaBuffer = await this.inner.outputSchema();
+    }
    const schema = tableFromIPC(schemaBuffer).schema;
    return schema;
  }
@@ -368,7 +356,7 @@ export class StandardQueryBase<
  extends QueryBase<NativeQueryType>
  implements ExecutableQuery
 {
-  constructor(inner?: NativeQueryType | Promise<NativeQueryType>) {
+  constructor(inner: NativeQueryType | Promise<NativeQueryType>) {
    super(inner);
  }

@@ -522,13 +510,6 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
    super(inner);
  }

-  /**
-   * @hidden
-   */
-  protected doVectorCall(fn: (inner: NativeVectorQuery) => void) {
-    super.doCall(fn);
-  }
-
  /**
   * Set the number of partitions to search (probe)
   *
@@ -556,7 +537,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * the minimum and maximum to the same value.
   */
  nprobes(nprobes: number): VectorQuery {
-    this.doVectorCall((inner) => inner.nprobes(nprobes));
+    super.doCall((inner) => inner.nprobes(nprobes));

    return this;
  }
@@ -570,7 +551,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * but will also increase latency.
   */
  minimumNprobes(minimumNprobes: number): VectorQuery {
-    this.doVectorCall((inner) => inner.minimumNprobes(minimumNprobes));
+    super.doCall((inner) => inner.minimumNprobes(minimumNprobes));
    return this;
  }

@@ -584,7 +565,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * potential false negatives.
   */
  maximumNprobes(maximumNprobes: number): VectorQuery {
-    this.doVectorCall((inner) => inner.maximumNprobes(maximumNprobes));
+    super.doCall((inner) => inner.maximumNprobes(maximumNprobes));
    return this;
  }

@@ -597,7 +578,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * `undefined` means no lower or upper bound.
   */
  distanceRange(lowerBound?: number, upperBound?: number): VectorQuery {
-    this.doVectorCall((inner) => inner.distanceRange(lowerBound, upperBound));
+    super.doCall((inner) => inner.distanceRange(lowerBound, upperBound));
    return this;
  }

@@ -611,7 +592,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * also increase the latency of your query. The default value is 1.5*limit.
   */
  ef(ef: number): VectorQuery {
-    this.doVectorCall((inner) => inner.ef(ef));
+    super.doCall((inner) => inner.ef(ef));
    return this;
  }

@@ -625,7 +606,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * whose data type is a fixed-size-list of floats.
   */
  column(column: string): VectorQuery {
-    this.doVectorCall((inner) => inner.column(column));
+    super.doCall((inner) => inner.column(column));
    return this;
  }

@@ -646,7 +627,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
  distanceType(
    distanceType: Required<IvfPqOptions>["distanceType"],
  ): VectorQuery {
-    this.doVectorCall((inner) => inner.distanceType(distanceType));
+    super.doCall((inner) => inner.distanceType(distanceType));
    return this;
  }

@@ -680,7 +661,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * distance between the query vector and the actual uncompressed vector.
   */
  refineFactor(refineFactor: number): VectorQuery {
-    this.doVectorCall((inner) => inner.refineFactor(refineFactor));
+    super.doCall((inner) => inner.refineFactor(refineFactor));
    return this;
  }

@@ -705,7 +686,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * factor can often help restore some of the results lost by post filtering.
   */
  postfilter(): VectorQuery {
-    this.doVectorCall((inner) => inner.postfilter());
+    super.doCall((inner) => inner.postfilter());
    return this;
  }

@@ -719,7 +700,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * calculate your recall to select an appropriate value for nprobes.
   */
  bypassVectorIndex(): VectorQuery {
-    this.doVectorCall((inner) => inner.bypassVectorIndex());
+    super.doCall((inner) => inner.bypassVectorIndex());
    return this;
  }

@@ -727,39 +708,43 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
   * Add a query vector to the search
   *
   * This method can be called multiple times to add multiple query vectors
-   * to the search. A column called `query_index` will be added to indicate the index
-   * of the query vector that produced the result. Flat searches share one table scan
-   * across the query vectors, avoiding the scan and memory amplification of running
-   * multiple queries concurrently. Indexed searches may still perform per-vector
-   * index work.
+   * to the search. If multiple query vectors are added, then they will be searched
+   * in parallel, and the results will be concatenated. A column called `query_index`
+   * will be added to indicate the index of the query vector that produced the result.
+   *
+   * Performance wise, this is equivalent to running multiple queries concurrently.
   */
  addQueryVector(vector: IntoVector): VectorQuery {
    if (vector instanceof Promise) {
-      // Observe the promise as soon as it is accepted. The existing native
-      // query may still be pending, and delaying observation until it resolves
-      // can otherwise surface a fast rejection as unhandled.
-      const settledVector = vector.then(
-        (value) => ({ status: "fulfilled" as const, value }),
-        (reason) => ({ status: "rejected" as const, reason }),
-      );
      const res = (async () => {
-        const inner = await this.getInner();
-        const outcome = await settledVector;
-        if (outcome.status === "rejected") {
-          throw outcome.reason;
+        try {
+          const v = await vector;
+          // biome-ignore lint/suspicious/noExplicitAny: we need to get the `inner`, but js has no package scoping
+          const value: any = this.addQueryVector(v);
+          const inner = value.inner as
+            | NativeVectorQuery
+            | Promise<NativeVectorQuery>;
+          return inner;
+        } catch (e) {
+          return Promise.reject(e);
        }
-        addQueryVectorToNative(inner, outcome.value);
-        return inner;
      })();
      return new VectorQuery(res);
    } else {
-      this.doVectorCall((inner) => addQueryVectorToNative(inner, vector));
+      super.doCall((inner) => {
+        const raw = Array.isArray(vector) ? null : extractVectorBuffer(vector);
+        if (raw) {
+          inner.addQueryVectorRaw(raw.data, raw.dtype);
+        } else {
+          inner.addQueryVector(Float32Array.from(vector as number[]));
+        }
+      });
      return this;
    }
  }

  rerank(reranker: Reranker): VectorQuery {
-    this.doVectorCall((inner) =>
+    super.doCall((inner) =>
      inner.rerank(async (args) => {
        const vecResults = await fromBufferToRecordBatch(args.vecResults);
        const ftsResults = await fromBufferToRecordBatch(args.ftsResults);
@@ -778,71 +763,6 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
  }
 }

-/**
- * Create a string query whose vector/FTS routing is resolved against the active
- * table schema when the query executes.
- *
- * @hidden
- */
-export function createAutoQuery(
-  table: NativeTable,
-  query: string,
-  columns: string[] | null,
-  getVector: (metadata: string) => Promise<Awaited<IntoVector>>,
-): AutoQuery {
-  type RouteSnapshot = {
-    table: NativeTable;
-    embeddingMetadata: string | undefined;
-  };
-  type CachedPreparation = {
-    metadata: string;
-    vector: Promise<Awaited<IntoVector>>;
-  };
-
-  let cachedPreparation: CachedPreparation | undefined;
-
-  const snapshotRoute = async (): Promise<RouteSnapshot> => {
-    const snapshot = await table.querySnapshot();
-    const schema = tableFromIPC(await snapshot.schema()).schema;
-    return {
-      table: snapshot,
-      embeddingMetadata: schema.metadata.get("embedding_functions"),
-    };
-  };
-
-  const createInner = async (): Promise<NativeQuery | NativeVectorQuery> => {
-    const route = await snapshotRoute();
-    if (route.embeddingMetadata === undefined) {
-      const inner = route.table.query();
-      inner.fullTextSearch({ query, columns });
-      return inner;
-    }
-
-    const metadata = route.embeddingMetadata;
-    if (cachedPreparation?.metadata !== metadata) {
-      cachedPreparation = {
-        metadata,
-        vector: Promise.resolve().then(() => getVector(metadata)),
-      };
-    }
-
-    const preparation = cachedPreparation;
-    let vector: Awaited<IntoVector>;
-    try {
-      vector = await preparation.vector;
-    } catch (error) {
-      if (cachedPreparation === preparation) {
-        cachedPreparation = undefined;
-      }
-      throw error;
-    }
-
-    return nearestToNative(route.table.query(), vector);
-  };
-
-  return new AutoQuery(createInner);
-}
-
 /**
 * A query that returns a subset of the rows in the table.
 *
@@ -868,51 +788,6 @@ export class TakeQuery extends QueryBase<NativeTakeQuery> {
  }
 }

-/**
- * A builder for automatic string searches.
- *
- * Automatic search determines whether to use full-text or vector search from
- * the table revision selected for each execution. This builder exposes the
- * common operations supported by both query families.
- *
- * @hideconstructor
- */
-export class AutoQuery extends StandardQueryBase<
-  NativeQuery | NativeVectorQuery
-> {
-  private readonly calls: Array<
-    (inner: NativeQuery | NativeVectorQuery) => void
-  > = [];
-
-  /** @hidden */
-  constructor(
-    private readonly createInner: () => Promise<
-      NativeQuery | NativeVectorQuery
-    >,
-  ) {
-    super();
-  }
-
-  /** @hidden */
-  protected override doCall(
-    fn: (inner: NativeQuery | NativeVectorQuery) => void,
-  ) {
-    this.calls.push(fn);
-  }
-
-  /** @hidden */
-  protected override async getInner(): Promise<
-    NativeQuery | NativeVectorQuery
-  > {
-    const calls = [...this.calls];
-    const inner = await this.createInner();
-    for (const call of calls) {
-      call(inner);
-    }
-    return inner;
-  }
-}
-
 /** A builder for LanceDB queries.
 *
 * @see {@link Table#query}, {@link Table#search}
@@ -965,19 +840,45 @@ export class Query extends StandardQueryBase<NativeQuery> {
   * a default `limit` of 10 will be used.  @see {@link Query#limit}
   */
  nearestTo(vector: IntoVector): VectorQuery {
-    const inner = this.inner;
-    if (inner instanceof Promise) {
-      const nativeQuery = inner.then(async (resolvedInner) =>
-        nearestToNative(resolvedInner, await vector),
-      );
+    const callNearestTo = (
+      inner: NativeQuery,
+      resolved: Float32Array | Float64Array | Uint8Array | number[],
+    ): NativeVectorQuery => {
+      const raw = Array.isArray(resolved)
+        ? null
+        : extractVectorBuffer(resolved);
+      if (raw) {
+        return inner.nearestToRaw(raw.data, raw.dtype);
+      }
+      return inner.nearestTo(Float32Array.from(resolved as number[]));
+    };
+
+    if (this.inner instanceof Promise) {
+      const nativeQuery = this.inner.then(async (inner) => {
+        const resolved = vector instanceof Promise ? await vector : vector;
+        return callNearestTo(inner, resolved);
+      });
      return new VectorQuery(nativeQuery);
    }
    if (vector instanceof Promise) {
-      return new VectorQuery(
-        vector.then((resolvedVector) => nearestToNative(inner, resolvedVector)),
-      );
+      const res = (async () => {
+        try {
+          const v = await vector;
+          // biome-ignore lint/suspicious/noExplicitAny: we need to get the `inner`, but js has no package scoping
+          const value: any = this.nearestTo(v);
+          const inner = value.inner as
+            | NativeVectorQuery
+            | Promise<NativeVectorQuery>;
+          return inner;
+        } catch (e) {
+          return Promise.reject(e);
+        }
+      })();
+      return new VectorQuery(res);
+    } else {
+      const vectorQuery = callNearestTo(this.inner, vector);
+      return new VectorQuery(vectorQuery);
    }
-    return new VectorQuery(nearestToNative(inner, vector));
  }

  nearestToText(query: string | FullTextQuery, columns?: string[]): Query {
@@ -94,24 +94,17 @@ export function sanitizeMetadata(
  if (metadataLike === undefined || metadataLike === null) {
    return undefined;
  }
-
-  let entries: IterableIterator<[unknown, unknown]>;
-  try {
-    entries = Map.prototype.entries.call(metadataLike);
-  } catch {
+  if (!(metadataLike instanceof Map)) {
    throw Error("Expected metadata, if present, to be a Map<string, string>");
  }
-
-  const metadata = new Map<string, string>();
-  for (const [key, value] of entries) {
-    if (typeof key !== "string" || typeof value !== "string") {
+  for (const item of metadataLike) {
+    if (typeof item[0] !== "string" || typeof item[1] !== "string") {
      throw Error(
        "Expected metadata, if present, to be a Map<string, string> but it had non-string keys or values",
      );
    }
-    metadata.set(key, value);
  }
-  return metadata;
+  return metadataLike as Map<string, string>;
 }

 export function sanitizeInt(typeLike: object) {
@@ -406,11 +406,10 @@ function matchingFields(fields: Field[], tree: FieldTree): Field[] {
          field.name,
          new Struct(matchingFields(struct.children, value)),
          field.nullable,
-          field.metadata,
        ),
      );
    } else {
-      matches.push(field);
+      matches.push(new Field(field.name, value as DataType, field.nullable));
    }
  }
  return matches;
@@ -43,12 +43,10 @@ import {
  Table as _NativeTable,
 } from "./native";
 import {
-  AutoQuery,
  FullTextQuery,
  Query,
  TakeQuery,
  VectorQuery,
-  createAutoQuery,
  instanceOfFullTextQuery,
 } from "./query";
 import { sanitizeType } from "./sanitize";
@@ -525,7 +523,7 @@ export abstract class Table {
    query: string | IntoVector | MultiVector | FullTextQuery,
    queryType?: string,
    ftsColumns?: string | string[],
-  ): VectorQuery | Query | AutoQuery;
+  ): VectorQuery | Query;
  /**
   * Search the table with a given query vector.
   *
@@ -630,18 +628,6 @@ export abstract class Table {

  /**
   * Update per-field (column) metadata.
-   *
-   * The following keys are treated specially, by convention, and should be
-   * used when appropriate:
-   *
-   * - `lancedb:description`: for a human-readable description of a field.
-   * - `lancedb:tag:<name>`: for a user-defined key-value tag, where the suffix
-   *   names the tag category; e.g. `lancedb:tag:model: "clip"`.
-   * - `lancedb:logical-column`: for a column grouping; e.g. `feature_v1` and
-   *   `feature_v2` might be in the same logical column.
-   * - `lancedb:status`: for status options (`production`, `candidate`,
-   *   `deprecated`, `archived`) to designate the current life cycle state of
-   *   this column.
   * @param {FieldMetadataUpdate[]} updates One or more per-field updates. Each
   * update's metadata is merged into the field's existing metadata by default;
   * a value of `null` deletes that key, and `replace: true` swaps the whole map.
@@ -989,11 +975,10 @@ export class LocalTable extends Table {
    return this.inner.display();
  }

-  private async getEmbeddingFunctions(
-    inner: _NativeTable = this.inner,
-  ): Promise<Map<string, EmbeddingFunctionConfig>> {
-    const schemaBuf = await inner.schema();
-    const schema = tableFromIPC(schemaBuf).schema;
+  private async getEmbeddingFunctions(): Promise<
+    Map<string, EmbeddingFunctionConfig>
+  > {
+    const schema = await this.schema();
    const registry = getRegistry();
    return registry.parseFunctions(schema.metadata);
  }
@@ -1175,7 +1160,7 @@ export class LocalTable extends Table {
    query: string | IntoVector | MultiVector | FullTextQuery,
    queryType: string = "auto",
    ftsColumns?: string | string[],
-  ): VectorQuery | Query | AutoQuery {
+  ): VectorQuery | Query {
    if (typeof query !== "string" && !instanceOfFullTextQuery(query)) {
      if (queryType === "fts") {
        throw new Error("Cannot perform full text search on a vector query");
@@ -1190,28 +1175,14 @@ export class LocalTable extends Table {
      });
    }

-    if (queryType === "auto") {
-      if (instanceOfFullTextQuery(query)) {
-        return this.query().fullTextSearch(query, {
-          columns: ftsColumns,
-        });
-      }
-
-      const columns =
-        typeof ftsColumns === "string" ? [ftsColumns] : (ftsColumns ?? null);
-      return createAutoQuery(this.inner, query, columns, async (metadata) => {
-        const functions = await getRegistry().parseFunctions(
-          new Map([["embedding_functions", metadata]]),
-        );
-        // TODO: Support multiple embedding functions
-        const embeddingFunc: EmbeddingFunctionConfig | undefined = functions
-          .values()
-          .next().value;
-        // The route only calls this callback when embedding metadata exists.
-        // parseFunctions either yields a provider or reports malformed metadata.
-        if (!embeddingFunc)
-          throw new Error("Invalid embedding function metadata");
-        return await embeddingFunc.function.computeQueryEmbeddings(query);
+    // The query type is auto or vector
+    // fall back to full text search if no embedding functions are defined and the query is a string
+    if (
+      queryType === "auto" &&
+      (getRegistry().length() === 0 || instanceOfFullTextQuery(query))
+    ) {
+      return this.query().fullTextSearch(query, {
+        columns: ftsColumns,
      });
    }

@@ -1567,8 +1538,7 @@ export interface FieldMetadataUpdate {
  path: string;
  /**
   * Metadata key/value pairs. Merged into the field's existing metadata by
-   * default; a value of `null` deletes that key. See
-   * {@link Table.updateFieldMetadata} for the conventional `lancedb:*` keys.
+   * default; a value of `null` deletes that key.
   */
  metadata: Record<string, string | null>;
  /** If true, replace the field's entire metadata map instead of merging. */
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",
@@ -1,6 +1,6 @@
 {
  "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.38.0-beta.12",
+  "version": "0.38.0-beta.10",
  "os": [
    "win32"
  ],
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.38.0-beta.12",
+	"version": "0.38.0-beta.10",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",
@@ -1,12 +1,12 @@
 {
  "name": "@lancedb/lancedb",
-  "version": "0.38.0-beta.12",
+  "version": "0.38.0-beta.9",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "@lancedb/lancedb",
-      "version": "0.38.0-beta.12",
+      "version": "0.38.0-beta.9",
      "cpu": [
        "x64",
        "arm64"
@@ -11,7 +11,7 @@
    "ann"
  ],
  "private": false,
-  "version": "0.38.0-beta.12",
+  "version": "0.38.0-beta.10",
  "main": "dist/index.js",
  "exports": {
    ".": "./dist/index.js",
@@ -278,13 +278,6 @@ impl Table {
        Ok(Query::new(self.inner_ref()?.query()))
    }

-    /// Return a read-only table handle pinned to the current query revision.
-    #[napi(catch_unwind)]
-    pub async fn query_snapshot(&self) -> napi::Result<Self> {
-        let snapshot = self.inner_ref()?.query_snapshot().await.default_error()?;
-        Ok(Self::new(snapshot))
-    }
-
    #[napi(catch_unwind)]
    pub fn take_offsets(&self, offsets: Vec<i64>) -> napi::Result<TakeQuery> {
        Ok(TakeQuery::new(
@@ -561,12 +554,6 @@ impl Table {
            .default_error()
    }

-    #[napi(catch_unwind)]
-    pub async fn checkout_current(&self) -> napi::Result<Self> {
-        let table = self.inner_ref()?.checkout_current().await.default_error()?;
-        Ok(Self::new(table))
-    }
-
    #[napi(catch_unwind)]
    pub async fn checkout(&self, version: i64) -> napi::Result<()> {
        self.inner_ref()?
@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.10"
 publish = false
 edition.workspace = true
 description = "Python bindings for LanceDB"
@@ -101,12 +101,9 @@ azure = ["adlfs>=2024.2.0"]
 [tool.maturin]
 python-source = "python"
 module-name = "lancedb._lancedb"
-# uv installs the project as an editable package before `uv run`, so keep that
-# bootstrap build consistent with `maturin develop`.
-editable-profile = "dev"

 [build-system]
-requires = ["maturin>=1.10"]
+requires = ["maturin>=1.9.4"]
 build-backend = "maturin"

 [tool.ruff.lint]
@@ -6,7 +6,7 @@ import importlib.metadata
 import os
 from concurrent.futures import ThreadPoolExecutor
 from datetime import timedelta
-from typing import Dict, Optional, Union, Any, List, Iterable, TYPE_CHECKING
+from typing import Dict, Optional, Union, Any, List, Iterable

 __version__ = importlib.metadata.version("lancedb")

@@ -20,7 +20,7 @@ from .db import AsyncConnection, DBConnection, LanceDBConnection
 from .remote import ClientConfig
 from .remote.db import RemoteDBConnection
 from .expr import Expr, col, lit, func
-from .schema import blob, vector
+from .schema import blob, vector, BlobType
 from .job import AsyncJob, Job
 from .functions import (
    FunctionArtifactRequest as FunctionArtifactRequest,
@@ -38,7 +38,7 @@ from .materialized_view import (
    MaterializedView,
    MaterializedViewDefinition,
 )
-from .table import AsyncTable, CompactionOptions, Table
+from .table import AsyncTable, Table
 from .types import BaseTokenizerType
 from ._lancedb import Session
 from .namespace import (
@@ -49,19 +49,6 @@ from .namespace import (
 )


-if TYPE_CHECKING:
-    from lance.blob import BlobType as BlobType
-
-
-def __getattr__(name: str):
-    if name == "BlobType":
-        from .schema import BlobType
-
-        globals()["BlobType"] = BlobType
-        return BlobType
-    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
-
-
 def _check_s3_bucket_with_dots(
    uri: str, storage_options: Optional[Dict[str, str]]
 ) -> None:
@@ -192,18 +179,6 @@ def connect(
    ...     },
    ... )

-    For Azure Blob Storage, credentials can be passed directly without setting
-    environment variables:
-
-    >>> azure_storage_options = {
-    ...     "account_name": "some-account",
-    ...     "account_key": "some-key",
-    ... }
-    >>> db = lancedb.connect(  # doctest: +SKIP
-    ...     "az://my-container/my-database",
-    ...     storage_options=azure_storage_options,
-    ... )
-
    For tests and temporary data, use an in-memory database:

    >>> db = lancedb.connect("memory://")
@@ -490,10 +465,6 @@ async def connect_async(
    --------

    >>> import lancedb
-    >>> azure_storage_options = {
-    ...     "account_name": "some-account",
-    ...     "account_key": "some-key",
-    ... }
    >>> async def doctest_example():
    ...     # For a local directory, provide a path to the database
    ...     db = await lancedb.connect_async("~/.lancedb")
@@ -501,11 +472,6 @@ async def connect_async(
    ...     db = await lancedb.connect_async("s3://my-bucket/lancedb",
    ...                                      storage_options={
    ...                                          "aws_access_key_id": "***"})
-    ...     # Azure credentials can also be passed directly
-    ...     db = await lancedb.connect_async(
-    ...         "az://my-container/my-database",
-    ...         storage_options=azure_storage_options,
-    ...     )
    ...     # For tests and temporary data, use an in-memory database
    ...     db = await lancedb.connect_async("memory://")
    ...     # Connect to LanceDB cloud
@@ -558,7 +524,6 @@ __all__ = [
    "AsyncJob",
    "AsyncLanceNamespaceDBConnection",
    "AsyncTable",
-    "CompactionOptions",
    "FtsToken",
    "col",
    "Expr",
@@ -12,7 +12,7 @@ from typing import TYPE_CHECKING, Optional, Union
 import pyarrow as pa

 from .expr import Expr
-from .schema import row_addressable_blob_v2_paths
+from .schema import blob_v2_column_paths
 from .types import BlobMode, QueryProjection, QueryProjectionSpec

 if TYPE_CHECKING:
@@ -119,7 +119,7 @@ def blob_v2_projection_sources(
    schema: pa.Schema,
    projection: QueryProjection,
 ) -> dict[str, str]:
-    blob_columns = row_addressable_blob_v2_paths(schema)
+    blob_columns = blob_v2_column_paths(schema)
    if not blob_columns:
        return {}
    columns = set(blob_columns)
@@ -140,9 +140,7 @@ def v2_projection_needs_row_id(
 ) -> bool:
    if with_row_id:
        return False
-    return projection_includes_blob_column(
-        projection, row_addressable_blob_v2_paths(schema)
-    )
+    return projection_includes_blob_column(projection, blob_v2_column_paths(schema))


 def blob_auto_row_id_for_scan(
@@ -272,8 +270,7 @@ def _iter_projection_pairs(
            if isinstance(expr, str):
                yield name, expr
            elif isinstance(expr, Expr):
-                source = expr._column_name()
-                yield name, source if source is not None else expr.to_sql()
+                yield name, expr.to_sql()
        return
    for column in projection:
        if isinstance(column, str):
@@ -283,8 +280,7 @@ def _iter_projection_pairs(
            if isinstance(expr, str):
                yield name, expr
            elif isinstance(expr, Expr):
-                source = expr._column_name()
-                yield name, source if source is not None else expr.to_sql()
+                yield name, expr.to_sql()


 def _set_blob_column(tbl: pa.Table, output_name: str, blobs: pa.Array) -> pa.Table:
@@ -87,7 +87,6 @@ class PyExpr:
    def contains(self, substr: "PyExpr") -> "PyExpr": ...
    def isin(self, values: List["PyExpr"]) -> "PyExpr": ...
    def cast(self, data_type: pa.DataType) -> "PyExpr": ...
-    def column_name(self) -> Optional[str]: ...
    def to_sql(self) -> str: ...

 def expr_col(name: str) -> PyExpr: ...
@@ -284,7 +283,6 @@ class Table:
        mode: Literal["append", "overwrite"],
        progress: Optional[Any] = None,
        write_parallelism: Optional[int] = None,
-        allow_external_blob_outside_bases: bool = False,
    ) -> AddResult: ...
    async def update(
        self, updates: Dict[str, str], where: Optional[str]
@@ -374,7 +372,6 @@ class Table:
        *,
        cleanup_since_ms: Optional[int] = None,
        delete_unverified: Optional[bool] = None,
-        compaction_options: Optional[Dict[str, Any]] = None,
    ) -> OptimizeStats: ...
    async def uri(self) -> str: ...
    async def initial_storage_options(self) -> Optional[Dict[str, str]]: ...
@@ -610,7 +607,6 @@ class PyQueryRequest:
    filter: Optional[Union[str, bytes]]
    full_text_search: Optional[FullTextQuery]
    select: Optional[Union[str, List[str]]]
-    select_source_columns: Optional[Dict[str, str]]
    fast_search: Optional[bool]
    with_row_id: Optional[bool]
    use_lsm: Optional[bool]
@@ -249,10 +249,6 @@ class Expr:

    # ── utilities ────────────────────────────────────────────────────────────

-    def _column_name(self) -> str | None:
-        """Return the source name when this is a bare column expression."""
-        return self._inner.column_name()
-
    def to_sql(self) -> str:
        """Render the expression as a SQL string (useful for debugging)."""
        return self._inner.to_sql()
@@ -316,7 +312,7 @@ def func(name: str, *args: ExprLike) -> Expr:
    --------
    >>> from lancedb.expr import col, func
    >>> func("lower", col("name"))
-    Expr(lower(`name`))
+    Expr(lower(name))
    """
    inner_args = [_coerce(a)._inner for a in args]
    return Expr(expr_func(name, inner_args))
@@ -222,7 +222,6 @@ class PythonEnvironmentSpec(_RemoteValue):

    kind: str
    packages: tuple[str, ...] = ()
-    channels: tuple[str, ...] = ()
    path: Optional[str] = None
    modules: tuple[str, ...] = ()
    image: Optional[str] = None
@@ -910,25 +909,13 @@ class UdfDefinition:
        pip: tuple[str, ...],
        env: Mapping[str, str],
        python_version: Optional[str],
-        conda: tuple[str, ...] = (),
-        conda_channels: tuple[str, ...] = (),
    ):
        function_name = name or function.__name__
        if not _FUNCTION_NAME.fullmatch(function_name):
            raise ValueError(f"invalid Function name: {function_name!r}")
-        if pip and conda:
-            raise ValueError("a Function environment is pip or conda, not both")
-        if conda_channels and not conda:
-            raise ValueError("conda_channels requires conda packages")
-        packages = tuple(sorted(set(conda if conda else pip)))
+        packages = tuple(sorted(set(pip)))
        if any(not package or package != package.strip() for package in packages):
-            raise ValueError("package requirements must be non-empty and trimmed")
-        if conda:
-            environment_spec = PythonEnvironmentSpec(
-                kind="conda", packages=packages, channels=tuple(conda_channels)
-            )
-        else:
-            environment_spec = PythonEnvironmentSpec(kind="pip", packages=packages)
+            raise ValueError("pip requirements must be non-empty and trimmed")
        environment = dict(env)
        if any(
            not isinstance(key, str) or not isinstance(value, str)
@@ -942,7 +929,7 @@ class UdfDefinition:
            kind="python",
            python_version=python_version
            or f"{sys.version_info.major}.{sys.version_info.minor}",
-            environment=environment_spec,
+            environment=PythonEnvironmentSpec(kind="pip", packages=packages),
            env=environment,
        )
        self._function = function
@@ -989,8 +976,6 @@ def udf(
    pip: tuple[str, ...] | list[str] = (),
    env: Optional[Mapping[str, str]] = None,
    python_version: Optional[str] = None,
-    conda: tuple[str, ...] | list[str] = (),
-    conda_channels: tuple[str, ...] | list[str] = (),
 ) -> Callable[[Callable[..., Any]], UdfDefinition]: ...


@@ -1003,8 +988,6 @@ def udf(
    pip: tuple[str, ...] | list[str] = (),
    env: Optional[Mapping[str, str]] = None,
    python_version: Optional[str] = None,
-    conda: tuple[str, ...] | list[str] = (),
-    conda_channels: tuple[str, ...] | list[str] = (),
 ):
    """Prepare a scalar Python callable for remote Function registration.

@@ -1027,10 +1010,6 @@ def udf(
        provided together with ``input_schema``.
    pip : sequence of str, optional
        Pip requirements for the remote environment.
-    conda : sequence of str, optional
-        Conda packages for the remote environment, instead of ``pip``.
-    conda_channels : sequence of str, optional
-        Conda channels in priority order; requires ``conda``.
    env : mapping of str to str, optional
        Environment variables included in the Function definition.
    python_version : str, optional
@@ -1070,8 +1049,6 @@ def udf(
            pip=tuple(pip),
            env={} if env is None else env,
            python_version=python_version,
-            conda=tuple(conda),
-            conda_channels=tuple(conda_channels),
        )

    if function is None:
@@ -7,7 +7,6 @@ from typing import List, Literal, Optional
 from ._lancedb import (
    IndexConfig,
 )
-from .query import DocumentGranularity
 from .types import BaseTokenizerType

 lang_mapping = {
@@ -122,11 +121,6 @@ class FTS:

    >>> config = FTS(block_size=256)

-    Create an index that treats each deepest-list element as one document:
-
-    >>> from lancedb.query import DocumentGranularity
-    >>> config = FTS(document_granularity=DocumentGranularity.LIST_ELEMENT)
-
    Attributes
    ----------
    with_position : bool, default False
@@ -178,11 +172,6 @@ class FTS:
        roughly half of the available CPU cores. The effective value is
        limited by the available compute capacity. This build-only setting is
        not persisted with the index and does not apply to remote tables.
-    document_granularity : DocumentGranularity, default ROW
-        ``ROW`` treats the selected text in one table row as one document.
-        ``LIST_ELEMENT`` treats each element of the deepest list on the indexed
-        field path as one document and returns its physical coordinates in
-        ``_doc_index`` for matching queries.

    Notes
    -----
@@ -207,7 +196,6 @@ class FTS:
    custom_stop_words: Optional[List[str]] = None
    memory_limit: Optional[int] = None
    num_workers: Optional[int] = None
-    document_granularity: DocumentGranularity = DocumentGranularity.ROW


@dataclass
@@ -167,12 +167,6 @@ def _projection_to_scanner_kwargs(columns: QueryProjection) -> Dict[str, Any]:
    return {"columns": projection}


-def _query_request_projection(req: "PyQueryRequest") -> QueryProjection:
-    if req.select_source_columns is not None:
-        return req.select_source_columns
-    return req.select
-
-
 def _scanner_kwargs_for_query(
    query: Query,
    blob_mode: BlobMode,
@@ -381,13 +375,6 @@ class FullTextOperator(str, Enum):
    OR = "OR"


-class DocumentGranularity(str, Enum):
-    """The unit treated as one full-text-search document."""
-
-    ROW = "row"
-    LIST_ELEMENT = "list_element"
-
-
 class Occur(str, Enum):
    SHOULD = "SHOULD"
    MUST = "MUST"
@@ -491,10 +478,6 @@ class MatchQuery(FullTextQuery):
    prefix_length : int, optional
        The number of beginning characters being unchanged for fuzzy matching.
        This is useful to achieve prefix matching.
-    document_granularity : DocumentGranularity, optional
-        Explicitly select row or deepest-list-element documents. If omitted,
-        the indexed granularity is inferred. When both granularities are indexed
-        for the field, this must be specified. With no index, row granularity is used.
    """

    query: str
@@ -504,9 +487,6 @@ class MatchQuery(FullTextQuery):
    max_expansions: int = pydantic.Field(50, kw_only=True)
    operator: FullTextOperator = pydantic.Field(FullTextOperator.OR, kw_only=True)
    prefix_length: int = pydantic.Field(0, kw_only=True)
-    document_granularity: Optional[DocumentGranularity] = pydantic.Field(
-        None, kw_only=True
-    )

    def query_type(self) -> FullTextQueryType:
        return FullTextQueryType.MATCH
@@ -523,20 +503,11 @@ class PhraseQuery(FullTextQuery):
        The query string to match against.
    column : str
        The name of the column to match against.
-    slop : int, default 0
-        The maximum number of intervening positions permitted in the phrase.
-    document_granularity : DocumentGranularity, optional
-        Explicitly select row or deepest-list-element documents. If omitted,
-        the indexed granularity is inferred. When both granularities are indexed
-        for the field, this must be specified. With no index, row granularity is used.
    """

    query: str
    column: str
    slop: int = pydantic.Field(0, kw_only=True)
-    document_granularity: Optional[DocumentGranularity] = pydantic.Field(
-        None, kw_only=True
-    )

    def query_type(self) -> FullTextQueryType:
        return FullTextQueryType.MATCH_PHRASE
@@ -2805,16 +2776,15 @@ class AsyncQueryBase(object):

        req = self._inner.to_query_request()
        schema = await self._table.schema()
-        projection = _query_request_projection(req)
        self._blob_auto_row_id = blob_auto_row_id_for_scan(
            schema,
-            projection,
+            req.select,
            with_row_id=self._with_row_id,
        )
        if not self._blob_auto_row_id:
            self._blob_paths = ()
            return
-        self._blob_paths = tuple(blob_v2_projection_sources(schema, projection).keys())
+        self._blob_paths = tuple(blob_v2_projection_sources(schema, req.select).keys())
        self._inner.with_row_id()

    def select(self, columns: Union[List[str], dict[str, str]]) -> Self:
@@ -3408,10 +3378,9 @@ class AsyncQuery(AsyncStandardQuery):
        pass in multiple vectors. When multiple vectors are passed in, if the vector
        column is with multivector type, then the vectors will be treated as a single
        query. Or the vectors will be treated as multiple queries, this can be useful
-        if you want to find the nearest vectors to multiple query vectors. Flat
-        searches share one table scan across the query vectors, avoiding the scan
-        and memory amplification of making multiple queries concurrently. If
-        multiple vectors are passed in then
+        if you want to find the nearest vectors to multiple query vectors.
+        This is not expected to be faster than making multiple queries concurrently;
+        it is just a convenience method. If multiple vectors are passed in then
        an additional column `query_index` will be added to the results. This column
        will contain the index of the query vector that the result is nearest to.
        """
@@ -3540,8 +3509,8 @@ class AsyncFTSQuery(AsyncStandardQuery):

        Typically, a single vector is passed in as the query. However, you can also
        pass in multiple vectors.  This can be useful if you want to find the nearest
-        vectors to multiple query vectors. Flat searches share one table scan across
-        the query vectors instead of issuing concurrent full scans.
+        vectors to multiple query vectors. This is not expected to be faster than
+        making multiple queries concurrently; it is just a convenience method.
        If multiple vectors are passed in then an additional column `query_index`
        will be added to the results.  This column will contain the index of the
        query vector that the result is nearest to.
@@ -3901,15 +3870,14 @@ class AsyncHybridQuery(AsyncStandardQuery, AsyncVectorQueryBase):
        blob_paths: tuple[str, ...] = ()
        if self._table is not None:
            schema = await self._table.schema()
-            projection = _query_request_projection(req)
            blob_auto_row_id = blob_auto_row_id_for_scan(
                schema,
-                projection,
+                req.select,
                with_row_id=self._with_row_id,
            )
            if blob_auto_row_id:
                blob_paths = tuple(
-                    blob_v2_projection_sources(schema, projection).keys()
+                    blob_v2_projection_sources(schema, req.select).keys()
                )
        self._blob_auto_row_id = blob_auto_row_id
        self._blob_paths = blob_paths
@@ -36,7 +36,6 @@ from lancedb._lancedb import (
    UpdateResult,
 )
 from lancedb.embeddings.base import EmbeddingFunctionConfig
-from lancedb.expr import Expr
 from lancedb.index import (
    FTS,
    BTree,
@@ -62,21 +61,11 @@ from lancedb.table import _normalize_progress

 from ..query import (
    AnalyzePlanDistributedMetrics,
-    DocumentGranularity,
    LanceQueryBuilder,
    LanceTakeQueryBuilder,
    LanceVectorQueryBuilder,
 )
-from ..table import (
-    AsyncTable,
-    BlobMode,
-    Branches,
-    CompactionOptions,
-    IndexStatistics,
-    Query,
-    Table,
-    Tags,
-)
+from ..table import AsyncTable, BlobMode, Branches, IndexStatistics, Query, Table, Tags
 from ..types import BaseTokenizerType


@@ -360,7 +349,6 @@ class RemoteTable(Table):
        ngram_max_length: int = 3,
        prefix_only: bool = False,
        block_size: int = 128,
-        document_granularity: DocumentGranularity = DocumentGranularity.ROW,
        name: Optional[str] = None,
    ):
        """Create a full-text search index on a column.
@@ -383,7 +371,6 @@ class RemoteTable(Table):
            ngram_max_length=ngram_max_length,
            prefix_only=prefix_only,
            block_size=block_size,
-            document_granularity=document_granularity,
        )
        LOOP.run(
            self._table.create_index(
@@ -623,7 +610,6 @@ class RemoteTable(Table):
        fill_value: float = 0.0,
        progress: Optional[Union[bool, Callable, Any]] = None,
        write_parallelism: Optional[int] = None,
-        allow_external_blob_outside_bases: bool = False,
    ) -> AddResult:
        """Add more data to the [Table][lancedb.table.Table].

@@ -656,8 +642,6 @@ class RemoteTable(Table):
            data in flight. Defaults to an estimate based on the data size,
            capped at the number of CPU cores. Lower this if bulk ingestion is
            using too much memory.
-        allow_external_blob_outside_bases: bool, default False
-            Not supported on LanceDB Cloud. Setting this raises.

        Returns
        -------
@@ -674,7 +658,6 @@ class RemoteTable(Table):
                    fill_value=fill_value,
                    progress=progress,
                    write_parallelism=write_parallelism,
-                    allow_external_blob_outside_bases=allow_external_blob_outside_bases,
                )
            )
        finally:
@@ -873,7 +856,7 @@ class RemoteTable(Table):

    def update(
        self,
-        where: Optional[Union[str, Expr]] = None,
+        where: Optional[str] = None,
        values: Optional[dict] = None,
        *,
        values_sql: Optional[Dict[str, str]] = None,
@@ -884,11 +867,9 @@ class RemoteTable(Table):

        Parameters
        ----------
-        where: str or [Expr][lancedb.expr.Expr], optional
-            The filter condition. Can be a SQL string or a type-safe
-            [Expr][lancedb.expr.Expr] built with [col][lancedb.expr.col] and
-            [lit][lancedb.expr.lit]. The filter must not be empty, or it will
-            error.
+        where: str, optional
+            The SQL where clause to use when updating rows. For example, 'x = 2'
+            or 'x IN (1, 2, 3)'. The filter must not be empty, or it will error.
        values: dict, optional
            The values to update. The keys are the column names and the values
            are the values to set.
@@ -962,7 +943,6 @@ class RemoteTable(Table):
        *,
        cleanup_older_than: Optional[timedelta] = None,
        delete_unverified: bool = False,
-        compaction_options: Optional[CompactionOptions] = None,
    ):
        """
        optimize() is a no-op on LanceDB Cloud.
@@ -4,34 +4,30 @@

 """Schema helpers for Lance blob columns."""

-import importlib
-from typing import TYPE_CHECKING
-
 import pyarrow as pa
-import pyarrow.ipc
-
-if TYPE_CHECKING:
-    from lance.blob import BlobType as BlobType

 _BLOB_EXTENSION_NAME = "lance.blob.v2"
 _BLOB_V1_KEY = "lance-encoding:blob"
 _ARROW_EXT_NAME_KEY = "ARROW:extension:name"
-_BLOB_V2_STORAGE_TYPE = pa.struct(
-    [
-        pa.field("data", pa.large_binary(), nullable=True),
-        pa.field("uri", pa.utf8(), nullable=True),
-        pa.field("position", pa.uint64(), nullable=True),
-        pa.field("size", pa.uint64(), nullable=True),
-    ]
-)
-_resolved_blob_type = None


-class _FallbackBlobType(pa.ExtensionType):
-    """lance.blob.v2 extension type used when pylance is not installed."""
+class BlobType(pa.ExtensionType):
+    """PyArrow extension type for a Lance blob v2 column.
+
+    Queries return descriptors; call :meth:`~lancedb.table.Table.fetch_blob_files`
+    for lazy reads or :meth:`~lancedb.table.Table.fetch_blobs` for eager bytes.
+    """

    def __init__(self) -> None:
-        pa.ExtensionType.__init__(self, _BLOB_V2_STORAGE_TYPE, _BLOB_EXTENSION_NAME)
+        storage_type = pa.struct(
+            [
+                pa.field("data", pa.large_binary(), nullable=True),
+                pa.field("uri", pa.utf8(), nullable=True),
+                pa.field("position", pa.uint64(), nullable=True),
+                pa.field("size", pa.uint64(), nullable=True),
+            ]
+        )
+        super().__init__(storage_type, _BLOB_EXTENSION_NAME)

    def __arrow_ext_serialize__(self) -> bytes:
        return b""
@@ -39,16 +35,23 @@ class _FallbackBlobType(pa.ExtensionType):
    @classmethod
    def __arrow_ext_deserialize__(
        cls, storage_type: pa.DataType, serialized: bytes
-    ) -> "_FallbackBlobType":
+    ) -> "BlobType":
        return cls()

    def __reduce__(self):
+        # Ensure pickle round-trips on older pyarrow (apache/arrow#35599).
        return type(self).__arrow_ext_deserialize__, (
            self.storage_type,
            self.__arrow_ext_serialize__(),
        )


+try:
+    pa.register_extension_type(BlobType())  # type: ignore[arg-type]
+except pa.ArrowKeyError:
+    pass
+
+
 def _metadata_value(metadata: dict, key: str):
    return metadata.get(key.encode()) or metadata.get(key)

@@ -89,105 +92,43 @@ def is_blob_like_field(field: pa.Field) -> bool:
    return is_blob_v2_field(field) or _metadata_marks_legacy_blob(field.metadata or {})


-def _collect_blob_paths(schema: pa.Schema, is_blob) -> list[tuple[str, bool]]:
-    """Walk the schema and return (path, has_list_ancestor) for each blob field."""
-    paths: list[tuple[str, bool]] = []
+def _collect_blob_paths(schema: pa.Schema, is_blob) -> list[str]:
+    paths: list[str] = []

-    def walk(fields, prefix: str, has_list_ancestor: bool) -> None:
+    def walk(fields, prefix: str) -> None:
        for field in fields:
            path = f"{prefix}.{field.name}" if prefix else field.name
            if is_blob(field):
-                paths.append((path, has_list_ancestor))
+                paths.append(path)
            elif pa.types.is_struct(field.type):
-                walk(field.type, path, has_list_ancestor)
+                walk(field.type, path)
            elif (
                pa.types.is_list(field.type)
                or pa.types.is_large_list(field.type)
                or pa.types.is_fixed_size_list(field.type)
            ):
-                walk([field.type.value_field], path, True)
+                walk([field.type.value_field], path)

-    walk(schema, "", False)
+    walk(schema, "")
    return paths


 def blob_column_paths(schema: pa.Schema) -> list[str]:
    """Dotted paths of blob-like columns (v2 extension or legacy metadata)."""
-    return [path for path, _ in _collect_blob_paths(schema, is_blob_like_field)]
+    return _collect_blob_paths(schema, is_blob_like_field)


 def blob_v2_column_paths(schema: pa.Schema) -> list[str]:
-    return [path for path, _ in _collect_blob_paths(schema, is_blob_v2_field)]
-
-
-def row_addressable_blob_v2_paths(schema: pa.Schema) -> list[str]:
-    """Blob v2 paths with one blob addressable by table row id.
-
-    ``fetch_blobs`` and the descriptor row-id ride-along address one blob per
-    row, so a blob inside a list container has no row-id slot and no fetch
-    path. Those columns still store and query as raw descriptors.
-    """
-    return [
-        path
-        for path, has_list_ancestor in _collect_blob_paths(schema, is_blob_v2_field)
-        if not has_list_ancestor
-    ]
+    return _collect_blob_paths(schema, is_blob_v2_field)


 def schema_has_blob_field(schema: pa.Schema) -> bool:
    return bool(blob_column_paths(schema))


-def _deserialize_registered_type(extension_type: pa.ExtensionType) -> pa.DataType:
-    """Return the type Arrow reconstructs for this extension name."""
-    schema = pa.schema([pa.field("value", extension_type)])
-    restored = pa.ipc.read_schema(schema.serialize())
-    return restored.field("value").type
-
-
-def _resolve_blob_type():
-    """Return the BlobType class this process should use.
-
-    pylance's class when it owns the lance.blob.v2 registry entry,
-    otherwise LanceDB's fallback. A different registered class is an error.
-    """
-    global _resolved_blob_type
-    if _resolved_blob_type is not None:
-        return _resolved_blob_type
-    try:
-        blob_module = importlib.import_module("lance.blob")
-    except ModuleNotFoundError as err:
-        if err.name not in ("lance", "lance.blob"):
-            raise
-    else:
-        blob_type = getattr(blob_module, "BlobType", None)
-        if blob_type is not None:
-            registered_type = _deserialize_registered_type(blob_type())
-            if type(registered_type) is not blob_type:
-                registered_cls = type(registered_type)
-                raise ValueError(
-                    "lance.blob.v2 is already registered by "
-                    f"{registered_cls.__module__}.{registered_cls.__qualname__}"
-                )
-            _resolved_blob_type = blob_type
-            return blob_type
-    try:
-        pa.register_extension_type(_FallbackBlobType())  # type: ignore[arg-type]
-    except pa.ArrowKeyError as err:
-        raise ValueError(
-            "lance.blob.v2 is already registered by another extension class"
-        ) from err
-    _resolved_blob_type = _FallbackBlobType
-    return _resolved_blob_type
-
-
 def blob(name: str, nullable: bool = True) -> pa.Field:
-    """Create a Lance blob v2 column field.
-
-    When pylance is installed this is ``lance.blob.BlobType``.
-    """
-    blob_type = _resolve_blob_type()
-    return pa.field(name, blob_type(), nullable=nullable)
+    """Create a Lance blob v2 column field."""
+    return pa.field(name, BlobType(), nullable=nullable)


 def vector(dimension: int, value_type: pa.DataType = pa.float32()) -> pa.DataType:
@@ -214,11 +155,3 @@ def vector(dimension: int, value_type: pa.DataType = pa.float32()) -> pa.DataTyp
    ... ])
    """
    return pa.list_(value_type, dimension)
-
-
-def __getattr__(name: str):
-    if name == "BlobType":
-        blob_type = _resolve_blob_type()
-        globals()["BlobType"] = blob_type
-        return blob_type
-    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
@@ -22,7 +22,6 @@ from typing import (
    Optional,
    Sequence,
    Tuple,
-    TypedDict,
    Union,
    overload,
 )
@@ -86,7 +85,6 @@ from .query import (
    AsyncQuery,
    AsyncTakeQuery,
    AsyncVectorQuery,
-    DocumentGranularity,
    FullTextQuery,
    LanceEmptyQueryBuilder,
    LanceFtsQueryBuilder,
@@ -105,12 +103,7 @@ from .util import (
    value_to_sql,
 )
 from .index import lang_mapping
-from .schema import (
-    blob_v2_column_paths,
-    is_blob_v2_field,
-    row_addressable_blob_v2_paths,
-    schema_has_blob_field,
-)
+from .schema import blob_v2_column_paths, schema_has_blob_field


 def _should_push_down_query_table(
@@ -228,88 +221,6 @@ IndexConfigType = Union[
    FTS,
 ]

-
-class CompactionOptions(TypedDict, total=False):
-    """Options that control file compaction during table optimization.
-
-    Unspecified options use Lance's defaults.
-
-    Compaction planning is row based. Lowering ``target_rows_per_fragment``
-    based on the expected row size can bound later compaction passes once
-    oversized fragments have been rewritten. It does not split an existing
-    fragment, so the first pass over an oversized fragment is not subject to
-    that bound. ``max_bytes_per_file`` limits output file size, not compaction
-    memory. Source budgets keep whole planned tasks; if the first task exceeds
-    a budget, that run performs no compaction work.
-
-    Examples
-    --------
-    Derive a steady-state row target from the expected row size:
-
-    >>> desired_fragment_bytes = 750 * 1024 * 1024
-    >>> average_row_bytes = 1_500_000
-    >>> options: CompactionOptions = {
-    ...     "target_rows_per_fragment": max(
-    ...         1, desired_fragment_bytes // average_row_bytes
-    ...     ),
-    ... }
-    >>> await table.optimize(compaction_options=options)  # doctest: +SKIP
-    """
-
-    target_rows_per_fragment: int
-    """Target rows per fragment; existing oversized fragments are not split."""
-
-    max_rows_per_group: int
-    """Maximum number of rows per row group (default: 1,024)."""
-
-    max_bytes_per_file: Optional[int]
-    """Maximum output data-file size; this does not bound compaction memory."""
-
-    materialize_deletions: bool
-    """Whether to rewrite fragments containing deleted rows (default: True)."""
-
-    materialize_deletions_threshold: float
-    """Minimum deleted-row fraction that makes a fragment eligible (default: 0.1)."""
-
-    num_threads: Optional[int]
-    """Number of compaction tasks to run in parallel."""
-
-    batch_size: Optional[int]
-    """Number of rows per input scan batch."""
-
-    io_buffer_size: Optional[int]
-    """Maximum number of bytes queued in the input scan I/O buffer."""
-
-    defer_index_remap: bool
-    """Whether to defer index remapping during compaction (default: False)."""
-
-    index_remap_mode: Literal["direct", "compact"]
-    """How to construct the old-to-new row-address mapping."""
-
-    compaction_mode: Optional[
-        Literal["reencode", "try_binary_copy", "force_binary_copy"]
-    ]
-    """Whether compaction re-encodes data or uses binary copying."""
-
-    binary_copy_read_batch_bytes: Optional[int]
-    """Number of bytes read per batch during binary-copy compaction."""
-
-    max_source_fragments: Optional[int]
-    """Maximum number of source fragments compacted in one run."""
-
-    max_source_rows: Optional[int]
-    """Maximum live source rows per run, applied to whole planned tasks."""
-
-    max_source_bytes: Optional[int]
-    """Maximum source bytes per run, applied to whole planned tasks."""
-
-    excluded_fragment_ids: List[int]
-    """Fragment IDs to leave unchanged and use as planning boundaries."""
-
-    max_overlays_per_fragment: Optional[int]
-    """Maximum overlays before a fragment is fully compacted."""
-
-
 # Known distance metrics for legacy API detection
 KNOWN_METRICS = {"l2", "cosine", "dot", "hamming"}

@@ -514,7 +425,6 @@ def _cast_to_target_schema(

    def gen():
        for batch in reader:
-            batch = _coerce_blob_write_columns(batch, reordered_schema)
            # Table but not RecordBatch has cast.
            cast_batches = (
                pa.Table.from_batches([batch]).cast(reordered_schema).to_batches()
@@ -527,166 +437,6 @@ def _cast_to_target_schema(
    return pa.RecordBatchReader.from_batches(reordered_schema, gen())


-def _coerce_blob_write_columns(
-    batch: pa.RecordBatch, target_schema: pa.Schema
-) -> pa.RecordBatch:
-    """Materialize blob storage structs before the stream leaves Python.
-
-    merge_insert requires its source reader to already match the table's
-    physical schema. Unlike add and insert, it does not pass through
-    LanceDB's Rust blob coercion, so preserving binary input here would
-    reach Lance as binary and fail the schema check.
-    """
-    columns = []
-    fields = []
-    changed = False
-    for field, column in zip(batch.schema, batch.columns):
-        target_field = target_schema.field(field.name)
-        coerced = _coerce_blob_value(column, target_field)
-        if coerced is not column:
-            column = coerced
-            field = pa.field(
-                field.name,
-                coerced.type,
-                field.nullable,
-                target_field.metadata,
-            )
-            changed = True
-        columns.append(column)
-        fields.append(field)
-    if not changed:
-        return batch
-    return pa.RecordBatch.from_arrays(
-        columns, schema=pa.schema(fields, metadata=batch.schema.metadata)
-    )
-
-
-def _coerce_blob_value(column: pa.Array, target_field: pa.Field) -> pa.Array:
-    if is_blob_v2_field(target_field) and _can_coerce_to_blob(column.type):
-        return _coerce_value_to_blob(column, target_field)
-
-    target_type = target_field.type
-    if pa.types.is_struct(target_type) and pa.types.is_struct(column.type):
-        children = []
-        fields = []
-        changed = False
-        for source_field in column.type:
-            source_column = column.field(source_field.name)
-            nested_target = next(
-                (field for field in target_type if field.name == source_field.name),
-                None,
-            )
-            if nested_target is None:
-                children.append(source_column)
-                fields.append(source_field)
-                continue
-            coerced = _coerce_blob_value(source_column, nested_target)
-            if coerced is not source_column:
-                changed = True
-            child_array, child_type = _physical_array_and_type(coerced)
-            children.append(child_array)
-            fields.append(
-                pa.field(
-                    source_field.name,
-                    child_type,
-                    source_field.nullable,
-                    nested_target.metadata,
-                )
-            )
-        if not changed:
-            return column
-        return pa.StructArray.from_arrays(
-            children,
-            fields=fields,
-            mask=column.is_null() if column.null_count else None,
-        )
-
-    if _is_list_like(target_type) and _is_list_like(column.type):
-        return _coerce_blob_list_values(column, target_type.value_field)
-
-    return column
-
-
-def _coerce_blob_list_values(
-    column: pa.Array, target_value_field: pa.Field
-) -> pa.Array:
-    """Coerce blob values inside a list column, preserving offsets and nulls.
-
-    Works on the raw child values window instead of ``pc.list_flatten`` because
-    flatten drops values spanned by null slots, which would misalign offsets.
-    """
-    mask = column.is_null() if column.null_count else None
-    if pa.types.is_fixed_size_list(column.type):
-        list_size = column.type.list_size
-        values = column.values.slice(column.offset * list_size, len(column) * list_size)
-        coerced = _coerce_blob_value(values, target_value_field)
-        if coerced is values:
-            return column
-        physical_values, _ = _physical_array_and_type(coerced)
-        return pa.FixedSizeListArray.from_arrays(physical_values, list_size, mask=mask)
-    offsets = column.offsets
-    first_offset = offsets[0].as_py()
-    values = column.values.slice(
-        first_offset,
-        offsets[-1].as_py() - first_offset,
-    )
-    coerced = _coerce_blob_value(values, target_value_field)
-    if coerced is values:
-        return column
-    physical_values, _ = _physical_array_and_type(coerced)
-    if first_offset:
-        offsets = pc.subtract(offsets, pa.scalar(first_offset, offsets.type))
-    if pa.types.is_large_list(column.type):
-        return pa.LargeListArray.from_arrays(offsets, physical_values, mask=mask)
-    return pa.ListArray.from_arrays(offsets, physical_values, mask=mask)
-
-
-def _coerce_value_to_blob(values: pa.Array, target_field: pa.Field) -> pa.Array:
-    if pa.types.is_null(values.type):
-        data = pa.nulls(len(values), type=pa.large_binary())
-    elif pa.types.is_large_binary(values.type):
-        data = values
-    else:
-        data = values.cast(pa.large_binary())
-    length = len(values)
-    storage_type = target_field.type
-    if isinstance(storage_type, pa.ExtensionType):
-        storage_type = storage_type.storage_type
-    storage_fields = list(storage_type)
-    children = []
-    for storage_field in storage_fields:
-        if storage_field.name == "data":
-            children.append(data)
-        else:
-            children.append(pa.nulls(length, type=storage_field.type))
-    storage = pa.StructArray.from_arrays(
-        children,
-        fields=storage_fields,
-        mask=values.is_null() if values.null_count else None,
-    )
-    if isinstance(target_field.type, pa.ExtensionType):
-        return pa.ExtensionArray.from_storage(target_field.type, storage)
-    return storage
-
-
-def _physical_array_and_type(array: pa.Array) -> tuple[pa.Array, pa.DataType]:
-    if isinstance(array.type, pa.ExtensionType):
-        return array.storage, array.type.storage_type
-    return array, array.type
-
-
-def _can_coerce_to_blob(data_type: pa.DataType) -> bool:
-    return _is_binary_like(data_type) or pa.types.is_null(data_type)
-
-
-def _is_binary_like(data_type: pa.DataType) -> bool:
-    return (
-        pa.types.is_binary(data_type)
-        or pa.types.is_large_binary(data_type)
-        or pa.types.is_binary_view(data_type)
-    )
-
-
 def _field_extension_name(field: pa.Field) -> Optional[str]:
    extension_name = getattr(field.type, "extension_name", None)
    if extension_name is not None:
@@ -713,73 +463,65 @@ def _align_field_types(
        target_field = next((f for f in target_fields if f.name == field.name), None)
        if target_field is None:
            raise ValueError(f"Field '{field.name}' not found in target schema")
-        new_fields.append(_align_field(field, target_field))
+        # Preserve arrow.json input until it reaches Lance. LanceDB exposes stored
+        # JSON columns as lance.json (JSONB-backed LargeBinary), but casting the
+        # input to that storage type here merely relabels the raw JSON bytes as
+        # JSONB. Lance must see arrow.json so it can perform the JSONB encoding.
+        if (
+            _field_extension_name(field) == "arrow.json"
+            and _field_extension_name(target_field) == "lance.json"
+        ):
+            new_fields.append(field)
+            continue
+        if pa.types.is_struct(target_field.type):
+            if pa.types.is_struct(field.type):
+                new_type = pa.struct(
+                    _align_field_types(
+                        field.type.fields,
+                        target_field.type.fields,
+                    )
+                )
+            else:
+                new_type = target_field.type
+        elif pa.types.is_list(target_field.type):
+            if _is_list_like(field.type):
+                new_type = pa.list_(
+                    _align_field_types(
+                        [field.type.value_field],
+                        [target_field.type.value_field],
+                    )[0]
+                )
+            else:
+                new_type = target_field.type
+        elif pa.types.is_large_list(target_field.type):
+            if _is_list_like(field.type):
+                new_type = pa.large_list(
+                    _align_field_types(
+                        [field.type.value_field],
+                        [target_field.type.value_field],
+                    )[0]
+                )
+            else:
+                new_type = target_field.type
+        elif pa.types.is_fixed_size_list(target_field.type):
+            if _is_list_like(field.type):
+                new_type = pa.list_(
+                    _align_field_types(
+                        [field.type.value_field],
+                        [target_field.type.value_field],
+                    )[0],
+                    target_field.type.list_size,
+                )
+            else:
+                new_type = target_field.type
+        else:
+            new_type = target_field.type
+        new_fields.append(
+            pa.field(field.name, new_type, field.nullable, target_field.metadata)
+        )
    return new_fields


-def _align_list_value_field(
-    value_field: pa.Field, target_value_field: pa.Field
-) -> pa.Field:
-    # A list has exactly one child, so the inferred child name ("item") aligns
-    # positionally and adopts the table's child name; pa.Table.cast renames it.
-    return _align_field(value_field, target_value_field).with_name(
-        target_value_field.name
-    )
-
-
-def _align_field(field: pa.Field, target_field: pa.Field) -> pa.Field:
-    # Preserve arrow.json input until it reaches Lance. LanceDB exposes stored
-    # JSON columns as lance.json (JSONB-backed LargeBinary), but casting the
-    # input to that storage type here merely relabels the raw JSON bytes as
-    # JSONB. Lance must see arrow.json so it can perform the JSONB encoding.
-    if (
-        _field_extension_name(field) == "arrow.json"
-        and _field_extension_name(target_field) == "lance.json"
-    ):
-        return field
-    if pa.types.is_struct(target_field.type):
-        if pa.types.is_struct(field.type):
-            new_type = pa.struct(
-                _align_field_types(
-                    field.type.fields,
-                    target_field.type.fields,
-                )
-            )
-        else:
-            new_type = target_field.type
-    elif pa.types.is_list(target_field.type):
-        if _is_list_like(field.type):
-            new_type = pa.list_(
-                _align_list_value_field(
-                    field.type.value_field, target_field.type.value_field
-                )
-            )
-        else:
-            new_type = target_field.type
-    elif pa.types.is_large_list(target_field.type):
-        if _is_list_like(field.type):
-            new_type = pa.large_list(
-                _align_list_value_field(
-                    field.type.value_field, target_field.type.value_field
-                )
-            )
-        else:
-            new_type = target_field.type
-    elif pa.types.is_fixed_size_list(target_field.type):
-        if _is_list_like(field.type):
-            new_type = pa.list_(
-                _align_list_value_field(
-                    field.type.value_field, target_field.type.value_field
-                ),
-                target_field.type.list_size,
-            )
-        else:
-            new_type = target_field.type
-    else:
-        new_type = target_field.type
-    return pa.field(field.name, new_type, field.nullable, target_field.metadata)
-
-
 def _infer_subschema(
    schema: List[pa.Field],
    reference_fields: List[pa.Field],
@@ -846,7 +588,7 @@ def sanitize_create_table(
        schema = data.schema
    else:
        if schema is not None:
-            data = pa.Table.from_batches([], schema=schema)
+            data = pa.Table.from_pylist([], schema)
    if schema is None:
        if data is None:
            raise ValueError("Either data or schema must be provided")
@@ -1426,7 +1168,6 @@ class Table(ABC):
        ngram_max_length: int = 3,
        prefix_only: bool = False,
        block_size: int = 128,
-        document_granularity: DocumentGranularity = DocumentGranularity.ROW,
        wait_timeout: Optional[timedelta] = None,
        name: Optional[str] = None,
    ):
@@ -1505,11 +1246,6 @@ class Table(ABC):
            The number of documents per compressed posting block. Must be 128
            or 256. A value of 256 uses the experimental FTS V3 format and
            may introduce breaking changes.
-        document_granularity: DocumentGranularity, default ROW
-            ``ROW`` treats the selected text in one table row as one document.
-            ``LIST_ELEMENT`` treats each element of the deepest list on the field
-            path as one document and returns its physical coordinates in
-            ``_doc_index`` for matching queries.
        wait_timeout: timedelta, optional
            The timeout to wait if indexing is asynchronous.
        name: str, optional
@@ -1533,7 +1269,6 @@ class Table(ABC):
        fill_value: float = 0.0,
        progress: Optional[Union[bool, Callable, Any]] = None,
        write_parallelism: Optional[int] = None,
-        allow_external_blob_outside_bases: bool = False,
    ) -> AddResult:
        """Add more data to the [Table][lancedb.table.Table].

@@ -1585,10 +1320,6 @@ class Table(ABC):
            data in flight. Defaults to an estimate based on the data size,
            capped at the number of CPU cores. Lower this if bulk ingestion is
            using too much memory.
-        allow_external_blob_outside_bases: bool, default False
-            Store blob URIs that sit outside registered blob bases. The row
-            keeps a reference, so the object has to stay readable. Local
-            tables only.

        Returns
        -------
@@ -2001,7 +1732,7 @@ class Table(ABC):
    @abstractmethod
    def update(
        self,
-        where: Optional[Union[str, Expr]] = None,
+        where: Optional[str] = None,
        values: Optional[dict] = None,
        *,
        values_sql: Optional[Dict[str, str]] = None,
@@ -2016,11 +1747,9 @@ class Table(ABC):

        Parameters
        ----------
-        where: str or [Expr][lancedb.expr.Expr], optional
-            The filter condition. Can be a SQL string or a type-safe
-            [Expr][lancedb.expr.Expr] built with [col][lancedb.expr.col] and
-            [lit][lancedb.expr.lit]. The filter must not be empty, or it will
-            error.
+        where: str, optional
+            The SQL where clause to use when updating rows. For example, 'x = 2'
+            or 'x IN (1, 2, 3)'. The filter must not be empty, or it will error.
        values: dict, optional
            The values to update. The keys are the column names and the values
            are the values to set.
@@ -2038,7 +1767,6 @@ class Table(ABC):
        Examples
        --------
        >>> import lancedb
-        >>> from lancedb.expr import col
        >>> import pandas as pd
        >>> data = pd.DataFrame({"x": [1, 2, 3], "vector": [[1.0, 2], [3, 4], [5, 6]]})
        >>> db = lancedb.connect("./.lancedb")
@@ -2048,7 +1776,7 @@ class Table(ABC):
        0  1  [1.0, 2.0]
        1  2  [3.0, 4.0]
        2  3  [5.0, 6.0]
-        >>> table.update(where=col("x") == 2, values={"vector": [10.0, 10]})
+        >>> table.update(where="x = 2", values={"vector": [10.0, 10]})
        UpdateResult(rows_updated=1, version=2)
        >>> table.to_pandas()
           x        vector
@@ -2132,7 +1860,6 @@ class Table(ABC):
        cleanup_older_than: Optional[timedelta] = None,
        delete_unverified: bool = False,
        retrain: bool = False,
-        compaction_options: Optional[CompactionOptions] = None,
    ):
        """
        Optimize the on-disk data and indices for better performance.
@@ -2165,11 +1892,6 @@ class Table(ABC):

        retrain: bool, default False
            This parameter is no longer used and is deprecated.
-        compaction_options: CompactionOptions, optional
-            Options that control file compaction. For large rows, derive a lower
-            ``target_rows_per_fragment`` from the expected row size to bound later
-            passes. This does not cap the first pass over an existing oversized
-            fragment; see [CompactionOptions][lancedb.table.CompactionOptions].

        Notes
        -----
@@ -2254,11 +1976,9 @@ class Table(ABC):
            Function columns are supported only on LanceDB Cloud and
            Enterprise.
        computed: Dict[str, str], optional
-            A mapping from output column names to SQL expressions derives each
-            output field from its expression. A direct projection of a Blob v2
-            field inherits Blob v2 semantics; other expressions derive their
-            ordinary Arrow type. Mapping order is declaration and dependency
-            order.
+            A map of column name to a SQL expression defining the column. The
+            column's type and inputs are derived from the expression, so no
+            data type is supplied.

            Unlike ``transforms``, the expression is stored rather than
            evaluated now: the column is committed with no values, and rows get
@@ -2395,25 +2115,12 @@ class Table(ABC):
        ----------
        updates : dict
            One or more dicts, each with:
-
            - "path": str — dot-path to the field (e.g. "embedding" or "a.b.c").
            - "metadata": dict[str, str | None] — keys to set; a value of ``None``
              deletes that key.
            - "replace": bool, optional — replace the field's whole metadata map
              instead of merging (default False).

-            The following keys are treated specially, by convention, and should
-            be used when appropriate:
-
-            - "lancedb:description": for a human-readable description of a field.
-            - ``"lancedb:tag:<name>"`` for a user-defined key-value tag, where the
-                suffix names the tag category; e.g. "lancedb:tag:model": "clip".
-            - "lancedb:logical-column" for a column grouping; e.g. "feature_v1"
-                and "feature_v2" might be in the same logical column.
-            - "lancedb:status" for status options ("production", "candidate",
-                "deprecated", "archived") to designate the current life cycle
-                state of this column.
-
        Returns
        -------
        UpdateFieldMetadataResult
@@ -2963,7 +2670,7 @@ class LanceTable(Table):
            arrow_tbl = self.to_arrow()
            if blob_mode == "descriptions":
                arrow_tbl = strip_auto_row_ids(
-                    arrow_tbl, row_addressable_blob_v2_paths(self.schema)
+                    arrow_tbl, blob_v2_column_paths(self.schema)
                )
            return arrow_tbl.to_pandas(**kwargs)

@@ -3561,7 +3268,6 @@ class LanceTable(Table):
        ngram_max_length: int = 3,
        prefix_only: bool = False,
        block_size: int = 128,
-        document_granularity: DocumentGranularity = DocumentGranularity.ROW,
        name: Optional[str] = None,
    ):
        """Create a full-text search index on a column.
@@ -3613,11 +3319,7 @@ class LanceTable(Table):
            tokenizer_configs = self.infer_tokenizer_configs(tokenizer_name)
            tokenizer_configs["custom_stop_words"] = custom_stop_words

-        config = FTS(
-            block_size=block_size,
-            document_granularity=document_granularity,
-            **tokenizer_configs,
-        )
+        config = FTS(block_size=block_size, **tokenizer_configs)

        try:
            LOOP.run(
@@ -3707,7 +3409,6 @@ class LanceTable(Table):
        fill_value: float = 0.0,
        progress: Optional[Union[bool, Callable, Any]] = None,
        write_parallelism: Optional[int] = None,
-        allow_external_blob_outside_bases: bool = False,
    ) -> AddResult:
        """Add data to the table.
        If vector columns are missing and the table
@@ -3735,9 +3436,6 @@ class LanceTable(Table):
            data in flight. Defaults to an estimate based on the data size,
            capped at the number of CPU cores. Lower this if bulk ingestion is
            using too much memory.
-        allow_external_blob_outside_bases: bool, default False
-            Allow blob URIs outside registered bases. See :meth:`Table.add`.
-            Local tables only.

        Returns
        -------
@@ -3754,7 +3452,6 @@ class LanceTable(Table):
                    fill_value=fill_value,
                    progress=progress,
                    write_parallelism=write_parallelism,
-                    allow_external_blob_outside_bases=allow_external_blob_outside_bases,
                )
            )
        finally:
@@ -4109,7 +3806,7 @@ class LanceTable(Table):

    def update(
        self,
-        where: Optional[Union[str, Expr]] = None,
+        where: Optional[str] = None,
        values: Optional[dict] = None,
        *,
        values_sql: Optional[Dict[str, str]] = None,
@@ -4120,11 +3817,9 @@ class LanceTable(Table):

        Parameters
        ----------
-        where: str or [Expr][lancedb.expr.Expr], optional
-            The filter condition. Can be a SQL string or a type-safe
-            [Expr][lancedb.expr.Expr] built with [col][lancedb.expr.col] and
-            [lit][lancedb.expr.lit]. The filter must not be empty, or it will
-            error.
+        where: str, optional
+            The SQL where clause to use when updating rows. For example, 'x = 2'
+            or 'x IN (1, 2, 3)'. The filter must not be empty, or it will error.
        values: dict, optional
            The values to update. The keys are the column names and the values
            are the values to set.
@@ -4142,7 +3837,6 @@ class LanceTable(Table):
        Examples
        --------
        >>> import lancedb
-        >>> from lancedb.expr import col
        >>> import pandas as pd
        >>> data = pd.DataFrame({"x": [1, 2, 3], "vector": [[1.0, 2], [3, 4], [5, 6]]})
        >>> db = lancedb.connect("./.lancedb")
@@ -4152,7 +3846,7 @@ class LanceTable(Table):
        0  1  [1.0, 2.0]
        1  2  [3.0, 4.0]
        2  3  [5.0, 6.0]
-        >>> table.update(where=col("x") == 2, values={"vector": [10.0, 10]})
+        >>> table.update(where="x = 2", values={"vector": [10.0, 10]})
        UpdateResult(rows_updated=1, version=2)
        >>> table.to_pandas()
           x        vector
@@ -4290,7 +3984,6 @@ class LanceTable(Table):
        cleanup_older_than: Optional[timedelta] = None,
        delete_unverified: bool = False,
        retrain: bool = False,
-        compaction_options: Optional[CompactionOptions] = None,
    ):
        """
        Optimize the on-disk data and indices for better performance.
@@ -4323,11 +4016,6 @@ class LanceTable(Table):

        retrain: bool, default False
            This parameter is no longer used and is deprecated.
-        compaction_options: CompactionOptions, optional
-            Options that control file compaction. For large rows, derive a lower
-            ``target_rows_per_fragment`` from the expected row size to bound later
-            passes. This does not cap the first pass over an existing oversized
-            fragment; see [CompactionOptions][lancedb.table.CompactionOptions].

        Notes
        -----
@@ -4342,7 +4030,6 @@ class LanceTable(Table):
                cleanup_older_than=cleanup_older_than,
                delete_unverified=delete_unverified,
                retrain=retrain,
-                compaction_options=compaction_options,
            )
        )

@@ -5374,9 +5061,7 @@ class AsyncTable:
        if blob_mode == "descriptions" or not schema_has_blob_field(schema):
            arrow_tbl = await self.to_arrow()
            if blob_mode == "descriptions":
-                arrow_tbl = strip_auto_row_ids(
-                    arrow_tbl, row_addressable_blob_v2_paths(schema)
-                )
+                arrow_tbl = strip_auto_row_ids(arrow_tbl, blob_v2_column_paths(schema))
            return arrow_tbl.to_pandas(**kwargs)

        if blob_mode == "lazy" and get_uri_scheme(await self.uri()) == "memory":
@@ -5680,7 +5365,6 @@ class AsyncTable:
        fill_value: Optional[float] = None,
        progress: Optional[Union[bool, Callable, Any]] = None,
        write_parallelism: Optional[int] = None,
-        allow_external_blob_outside_bases: bool = False,
    ) -> AddResult:
        """Add more data to the [AsyncTable][lancedb.table.AsyncTable].

@@ -5711,9 +5395,6 @@ class AsyncTable:
            data in flight. Defaults to an estimate based on the data size,
            capped at the number of CPU cores. Lower this if bulk ingestion is
            using too much memory.
-        allow_external_blob_outside_bases: bool, default False
-            Allow blob URIs outside registered bases. See :meth:`Table.add`.
-            Local tables only.

        """
        schema = await self.schema()
@@ -5750,7 +5431,6 @@ class AsyncTable:
                mode or "append",
                progress=progress,
                write_parallelism=write_parallelism,
-                allow_external_blob_outside_bases=allow_external_blob_outside_bases,
            )
        except RuntimeError as e:
            if "Cast error" in str(e):
@@ -6275,7 +5955,7 @@ class AsyncTable:
        self,
        updates: Optional[Dict[str, Any]] = None,
        *,
-        where: Optional[Union[str, Expr]] = None,
+        where: Optional[str] = None,
        updates_sql: Optional[Dict[str, str]] = None,
    ) -> UpdateResult:
        """
@@ -6290,11 +5970,9 @@ class AsyncTable:
            The updates to apply.  The keys should be the name of the column to
            update.  The values should be the new values to assign.  This is
            required unless updates_sql is supplied.
-        where: str or [Expr][lancedb.expr.Expr], optional
-            The filter condition. Can be a SQL string or a type-safe
-            [Expr][lancedb.expr.Expr] built with [col][lancedb.expr.col] and
-            [lit][lancedb.expr.lit]. Only rows that satisfy this filter will
-            be updated.
+        where: str, optional
+            An SQL filter that controls which rows are updated. For example, 'x = 2'
+            or 'x IN (1, 2, 3)'.  Only rows that satisfy this filter will be udpated.
        updates_sql: dict, optional
            The updates to apply, expressed as SQL expression strings.  The keys should
            be column names. The values should be SQL expressions.  These can be SQL
@@ -6312,14 +5990,13 @@ class AsyncTable:
        --------
        >>> import asyncio
        >>> import lancedb
-        >>> from lancedb.expr import col
        >>> import pandas as pd
        >>> async def demo_update():
        ...     data = pd.DataFrame({"x": [1, 2], "vector": [[1, 2], [3, 4]]})
        ...     db = await lancedb.connect_async("./.lancedb")
        ...     table = await db.create_table("my_table", data)
        ...     # x is [1, 2], vector is [[1, 2], [3, 4]]
-        ...     await table.update({"vector": [10, 10]}, where=col("x") == 2)
+        ...     await table.update({"vector": [10, 10]}, where="x = 2")
        ...     # x is [1, 2], vector is [[1, 2], [10, 10]]
        ...     await table.update(updates_sql={"x": "x + 1"})
        ...     # x is [2, 3], vector is [[1, 2], [10, 10]]
@@ -6333,8 +6010,7 @@ class AsyncTable:
        if updates is not None:
            updates_sql = {k: value_to_sql(v) for k, v in updates.items()}

-        predicate = where.to_sql() if isinstance(where, Expr) else where
-        return await self._inner.update(updates_sql, predicate)
+        return await self._inner.update(updates_sql, where)

    async def add_columns(
        self,
@@ -6366,11 +6042,8 @@ class AsyncTable:
            Function columns are supported only on LanceDB Cloud and
            Enterprise.
        computed: Dict[str, str], optional
-            A mapping from output column names to SQL expressions derives each
-            output field from its expression. A direct projection of a Blob v2
-            field inherits Blob v2 semantics; other expressions derive their
-            ordinary Arrow type. Mapping order is declaration and dependency
-            order.
+            A map of column name to a SQL expression defining the column. The
+            column's type and inputs are derived from the expression.

            Unlike ``transforms``, the expression is stored rather than
            evaluated now: the column is committed with no values, and rows get
@@ -6744,7 +6417,6 @@ class AsyncTable:
        cleanup_older_than: Optional[timedelta] = None,
        delete_unverified: bool = False,
        retrain=False,
-        compaction_options: Optional[CompactionOptions] = None,
    ) -> OptimizeStats:
        """
        Optimize the on-disk data and indices for better performance.
@@ -6777,11 +6449,6 @@ class AsyncTable:

        retrain: bool, default False
            This parameter is no longer used and is deprecated.
-        compaction_options: CompactionOptions, optional
-            Options that control file compaction. For large rows, derive a lower
-            ``target_rows_per_fragment`` from the expected row size to bound later
-            passes. This does not cap the first pass over an existing oversized
-            fragment; see [CompactionOptions][lancedb.table.CompactionOptions].

        Notes
        -----
@@ -6808,7 +6475,6 @@ class AsyncTable:
        return await self._inner.optimize(
            cleanup_since_ms=cleanup_since_ms,
            delete_unverified=delete_unverified,
-            compaction_options=compaction_options,
        )

    async def list_indices(self) -> Iterable[IndexConfig]:
@@ -105,7 +105,7 @@ def test_quickstart(tmp_path):
    tbl.create_index(num_sub_vectors=1)
    # --8<-- [end:create_index]
    # --8<-- [start:delete_rows]
-    tbl.delete("item = 'fizz'")
+    tbl.delete('item = "fizz"')
    # --8<-- [end:delete_rows]
    # --8<-- [start:drop_table]
    db.drop_table("my_table")
@@ -201,7 +201,7 @@ async def test_quickstart_async(tmp_path):
    await tbl.create_index("vector")
    # --8<-- [end:create_index_async]
    # --8<-- [start:delete_rows_async]
-    await tbl.delete("item = 'fizz'")
+    await tbl.delete('item = "fizz"')
    # --8<-- [end:delete_rows_async]
    # --8<-- [start:drop_table_async]
    await db.drop_table("my_table_async")
@@ -266,7 +266,7 @@ def test_table():
    tbl.add(pydantic_model_items)
    # --8<-- [end:add_table_from_pydantic]
    # --8<-- [start:delete_row]
-    tbl.delete("item = 'fizz'")
+    tbl.delete('item = "fizz"')
    # --8<-- [end:delete_row]
    # --8<-- [start:delete_specific_row]
    data = [
@@ -538,7 +538,7 @@ async def test_table_async():
    await async_tbl.add(pydantic_model_items)
    # --8<-- [end:add_table_async_from_pydantic]
    # --8<-- [start:delete_row_async]
-    await async_tbl.delete("item = 'fizz'")
+    await async_tbl.delete('item = "fizz"')
    # --8<-- [end:delete_row_async]
    # --8<-- [start:delete_specific_row_async]
    data = [
@@ -2,41 +2,17 @@
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors

 import io
-import subprocess
-import sys
-import textwrap

-import lance
 import pyarrow as pa
 import pyarrow.compute as pc
 import pytest
-from lance.blob import BlobType as LanceBlobType

 import lancedb
-from lancedb._blob import (
-    blob_v2_projection_sources,
-    read_row_ids_from_hits,
-    stash_auto_row_ids,
-)
-from lancedb.expr import col
+from lancedb._blob import read_row_ids_from_hits, stash_auto_row_ids
 from lancedb.index import FTS
 from lancedb.schema import blob_column_paths, blob_v2_column_paths


-_HIDE_LANCE_BLOB = """\
-import importlib.abc
-import sys
-
-class _MissingLanceBlob(importlib.abc.MetaPathFinder):
-    def find_spec(self, fullname, path, target=None):
-        if fullname == "lance.blob" or fullname.startswith("lance.blob."):
-            raise ModuleNotFoundError(fullname, name="lance.blob")
-
-sys.modules.pop("lance.blob", None)
-sys.meta_path.insert(0, _MissingLanceBlob())
-"""
-
-
 def _blob_table(name, rows):
    db = lancedb.connect("memory:///")
    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
@@ -70,181 +46,6 @@ def test_blob_factory_declares_v2_field():
    field = lancedb.blob("image")
    assert isinstance(field.type, pa.ExtensionType)
    assert field.type.extension_name == "lance.blob.v2"
-    assert lancedb.BlobType is LanceBlobType
-    assert type(field.type) is LanceBlobType
-
-
-def test_blob_type_works_without_pylance():
-    script = _HIDE_LANCE_BLOB + textwrap.dedent(
-        """\
-        import lancedb
-        import pyarrow as pa
-
-        field = lancedb.blob("image")
-        if not isinstance(field.type, pa.ExtensionType):
-            raise SystemExit("expected an extension type")
-        if field.type.extension_name != "lance.blob.v2":
-            raise SystemExit(field.type.extension_name)
-        if lancedb.BlobType is not type(field.type):
-            raise SystemExit("BlobType is not the field type class")
-        if lancedb.BlobType.__module__ != "lancedb.schema":
-            raise SystemExit(lancedb.BlobType.__module__)
-
-        db = lancedb.connect("memory:///")
-        table = db.create_table(
-            "images",
-            schema=pa.schema([pa.field("id", pa.int64()), field]),
-        )
-        table.add([{"id": 1, "image": b"hello"}])
-        result = (
-            table.merge_insert("id")
-            .when_matched_update_all()
-            .when_not_matched_insert_all()
-            .execute([{"id": 1, "image": b"updated"}, {"id": 2, "image": b"inserted"}])
-        )
-        if result.num_updated_rows != 1 or result.num_inserted_rows != 1:
-            raise SystemExit(
-                f"merge_insert rows updated={result.num_updated_rows} "
-                f"inserted={result.num_inserted_rows}"
-            )
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_blob_resolves_pylance_type_without_eager_import():
-    script = textwrap.dedent(
-        """\
-        import sys
-        import lancedb
-
-        if "lance.blob" in sys.modules:
-            raise SystemExit("import lancedb imported lance.blob")
-        field = lancedb.blob("image")
-        from lance.blob import BlobType
-
-        if type(field.type) is not BlobType:
-            raise SystemExit(f"{type(field.type)} is not {BlobType}")
-        import lance
-
-        image = lance.blob_array([b"x"])
-        if type(image.type) is not BlobType:
-            raise SystemExit("blob_array used a different class")
-        if type(image.type) is not type(field.type):
-            raise SystemExit("field and array classes differ")
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_blob_fallback_fails_if_name_already_registered():
-    script = _HIDE_LANCE_BLOB + textwrap.dedent(
-        """\
-        import pyarrow as pa
-
-        class OtherBlobType(pa.ExtensionType):
-            def __init__(self):
-                super().__init__(
-                    pa.struct([pa.field("data", pa.large_binary())]),
-                    "lance.blob.v2",
-                )
-
-            def __arrow_ext_serialize__(self):
-                return b""
-
-            @classmethod
-            def __arrow_ext_deserialize__(cls, storage_type, serialized):
-                return cls()
-
-        pa.register_extension_type(OtherBlobType())
-        import lancedb
-
-        try:
-            lancedb.blob("image")
-        except ValueError as err:
-            if "already registered" not in str(err):
-                raise SystemExit(err)
-        else:
-            raise SystemExit("expected ValueError")
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_blob_type_rejects_competing_registration_with_pylance():
-    script = textwrap.dedent(
-        """\
-        import pyarrow as pa
-        import pyarrow.ipc
-
-        class OtherBlobType(pa.ExtensionType):
-            def __init__(self):
-                super().__init__(
-                    pa.struct(
-                        [
-                            pa.field("data", pa.large_binary()),
-                            pa.field("uri", pa.utf8()),
-                            pa.field("position", pa.uint64()),
-                            pa.field("size", pa.uint64()),
-                        ]
-                    ),
-                    "lance.blob.v2",
-                )
-
-            def __arrow_ext_serialize__(self):
-                return b""
-
-            @classmethod
-            def __arrow_ext_deserialize__(cls, storage_type, serialized):
-                return cls()
-
-        pa.register_extension_type(OtherBlobType())
-
-        from lance.blob import BlobType
-
-        if BlobType is OtherBlobType:
-            raise SystemExit("pylance BlobType was replaced")
-        schema = pa.schema([pa.field("value", BlobType())])
-        restored = pa.ipc.read_schema(schema.serialize())
-        if type(restored.field("value").type) is not OtherBlobType:
-            raise SystemExit(type(restored.field("value").type))
-
-        import lancedb
-
-        try:
-            lancedb.blob("image")
-        except ValueError as err:
-            if "__main__.OtherBlobType" not in str(err):
-                raise SystemExit(err)
-        else:
-            raise SystemExit("expected ValueError")
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr


 def test_blob_v2_column_paths_include_list_children():
@@ -269,14 +70,6 @@ def test_blob_v2_column_paths_include_list_children():
    ]


-def test_blob_v2_projection_sources_use_typed_column_name():
-    schema = pa.schema([lancedb.blob("blob")])
-
-    assert blob_v2_projection_sources(schema, {"blob_alias": col("blob")}) == {
-        "blob_alias": "blob"
-    }
-
-
 def _legacy_v1_table(name):
    db = lancedb.connect("memory:///")
    schema = pa.schema(
@@ -373,20 +166,6 @@ async def test_async_table_to_pandas_descriptions_mode_omits_row_id():
    assert set(descriptor.keys()) == {"kind", "position", "size", "blob_id", "blob_uri"}


-@pytest.mark.asyncio
-async def test_async_typed_blob_projection_preserves_source_column():
-    db = await lancedb.connect_async("memory:///typed_blob_projection")
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("blob")])
-    table = await db.create_table("typed_blob_projection", schema=schema)
-    await table.add([{"id": 1, "blob": b"alpha"}])
-
-    hits = await table.query().select({"blob_alias": col("blob")}).to_arrow()
-
-    assert "_lance_row_id" in hits.schema.field("blob_alias").type.names
-    blobs = await table.fetch_blobs("blob", hits)
-    assert blobs.to_pylist() == [b"alpha"]
-
-
 def test_fetch_blobs_round_trip():
    table = _blob_table(
        "round_trip",
@@ -397,292 +176,6 @@ def test_fetch_blobs_round_trip():
    assert [blobs[0].as_py(), blobs[1].as_py()] == [b"alpha", b"beta"]


-def test_merge_insert_writes_python_bytes():
-    table = _blob_table("merge_bytes", [{"id": 1, "image": b"before"}])
-    result = (
-        table.merge_insert("id")
-        .when_matched_update_all()
-        .when_not_matched_insert_all()
-        .execute([{"id": 1, "image": b"updated"}, {"id": 2, "image": b"inserted"}])
-    )
-    assert result.num_updated_rows == 1
-    assert result.num_inserted_rows == 1
-    by_id = _row_ids_by_id(table)
-    blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-    assert blobs.to_pylist() == [b"updated", b"inserted"]
-
-
-def test_merge_insert_bytes_after_reopen_without_touching_blob_type(tmp_path):
-    db = lancedb.connect(tmp_path)
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table("images", schema=schema)
-    table.add([{"id": 1, "image": b"hello"}])
-
-    script = textwrap.dedent(
-        f"""\
-        import lancedb
-
-        db = lancedb.connect({str(tmp_path)!r})
-        table = db.open_table("images")
-        image_type = table.schema.field("image").type
-        if type(image_type).__name__ != "StructType":
-            raise SystemExit(f"expected StructType, got {{type(image_type)}}")
-        result = (
-            table.merge_insert("id")
-            .when_matched_update_all()
-            .when_not_matched_insert_all()
-            .execute(
-                [{{"id": 1, "image": b"updated"}}, {{"id": 2, "image": b"inserted"}}]
-            )
-        )
-        if result.num_updated_rows != 1 or result.num_inserted_rows != 1:
-            raise SystemExit(
-                f"rows updated={{result.num_updated_rows}} "
-                f"inserted={{result.num_inserted_rows}}"
-            )
-        hits = table.search().with_row_id(True).limit(10).to_arrow()
-        by_id = dict(zip(hits["id"].to_pylist(), hits["_rowid"].to_pylist()))
-        blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-        if blobs.to_pylist() != [b"updated", b"inserted"]:
-            raise SystemExit(blobs.to_pylist())
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_merge_insert_bytes_after_reopen_without_pylance(tmp_path):
-    db = lancedb.connect(tmp_path)
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table("images", schema=schema)
-    table.add([{"id": 1, "image": b"hello"}])
-
-    script = _HIDE_LANCE_BLOB + textwrap.dedent(
-        f"""\
-        import lancedb
-
-        db = lancedb.connect({str(tmp_path)!r})
-        table = db.open_table("images")
-        image_type = table.schema.field("image").type
-        if type(image_type).__name__ != "StructType":
-            raise SystemExit(f"expected StructType, got {{type(image_type)}}")
-        result = (
-            table.merge_insert("id")
-            .when_matched_update_all()
-            .when_not_matched_insert_all()
-            .execute(
-                [{{"id": 1, "image": b"updated"}}, {{"id": 2, "image": b"inserted"}}]
-            )
-        )
-        if result.num_updated_rows != 1 or result.num_inserted_rows != 1:
-            raise SystemExit(
-                f"rows updated={{result.num_updated_rows}} "
-                f"inserted={{result.num_inserted_rows}}"
-            )
-        hits = table.search().with_row_id(True).limit(10).to_arrow()
-        by_id = dict(zip(hits["id"].to_pylist(), hits["_rowid"].to_pylist()))
-        blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-        if blobs.to_pylist() != [b"updated", b"inserted"]:
-            raise SystemExit(blobs.to_pylist())
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_merge_insert_blob_array_into_reopened_unregistered_table(tmp_path):
-    db = lancedb.connect(tmp_path)
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table("images", schema=schema)
-    table.add([{"id": 1, "image": b"before"}])
-
-    script = textwrap.dedent(
-        f"""\
-        import pyarrow as pa
-        import lancedb
-
-        db = lancedb.connect({str(tmp_path)!r})
-        table = db.open_table("images")
-        image_type = table.schema.field("image").type
-        if type(image_type).__name__ != "StructType":
-            raise SystemExit(
-                f"expected StructType before lance import, got {{type(image_type)}}"
-            )
-
-        import lance
-
-        updates = pa.Table.from_arrays(
-            [
-                pa.array([1, 2], type=pa.int64()),
-                lance.blob_array([b"updated", b"inserted"]),
-            ],
-            names=["id", "image"],
-        )
-        result = (
-            table.merge_insert("id")
-            .when_matched_update_all()
-            .when_not_matched_insert_all()
-            .execute(updates)
-        )
-        if result.num_updated_rows != 1 or result.num_inserted_rows != 1:
-            raise SystemExit(
-                f"rows updated={{result.num_updated_rows}} "
-                f"inserted={{result.num_inserted_rows}}"
-            )
-        hits = table.search().with_row_id(True).limit(10).to_arrow()
-        by_id = dict(zip(hits["id"].to_pylist(), hits["_rowid"].to_pylist()))
-        blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-        if blobs.to_pylist() != [b"updated", b"inserted"]:
-            raise SystemExit(blobs.to_pylist())
-        """
-    )
-    result = subprocess.run(
-        [sys.executable, "-c", script],
-        capture_output=True,
-        text=True,
-        check=False,
-    )
-    assert result.returncode == 0, result.stderr
-
-
-def test_add_all_null_blob_column():
-    db = lancedb.connect("memory:///")
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table("all_null", schema=schema)
-    table.add([{"id": 1, "image": None}, {"id": 2, "image": None}])
-    by_id = _row_ids_by_id(table)
-    blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-    assert blobs.to_pylist() == [None, None]
-
-
-def test_create_table_nested_blob_schema_without_rows():
-    db = lancedb.connect("memory:///")
-    schema = pa.schema(
-        [
-            pa.field("id", pa.int64()),
-            pa.field("info", pa.struct([lancedb.blob("blob")])),
-            pa.field("images", pa.list_(lancedb.blob("image"))),
-        ]
-    )
-    table = db.create_table("nested_empty", schema=schema)
-    assert table.count_rows() == 0
-
-
-def test_merge_insert_nested_blob_dicts():
-    db = lancedb.connect("memory:///")
-    info = pa.StructArray.from_arrays(
-        [
-            pa.array(["first"], type=pa.string()),
-            _blob_array("blob", [b"before"]),
-        ],
-        names=["name", "blob"],
-    )
-    data = pa.Table.from_arrays(
-        [pa.array([1], type=pa.int64()), info],
-        names=["id", "info"],
-    )
-    table = db.create_table("nested_merge", data=data)
-    result = (
-        table.merge_insert("id")
-        .when_matched_update_all()
-        .execute([{"id": 1, "info": {"name": "first", "blob": b"after"}}])
-    )
-    assert result.num_updated_rows == 1
-    by_id = _row_ids_by_id(table)
-    blobs = table.fetch_blobs("info.blob", [by_id[1]])
-    assert blobs.to_pylist() == [b"after"]
-
-
-def _list_blob_table(name):
-    db = lancedb.connect("memory:///")
-    blob_field = lancedb.blob("image")
-    images = pa.ListArray.from_arrays(
-        pa.array([0, 1], type=pa.int32()), _blob_array("image", [b"before"])
-    )
-    data = pa.Table.from_arrays(
-        [pa.array([1], type=pa.int64()), images],
-        schema=pa.schema(
-            [pa.field("id", pa.int64()), pa.field("images", pa.list_(blob_field))]
-        ),
-    )
-    return db.create_table(name, data=data)
-
-
-def test_merge_insert_list_blob_dicts():
-    table = _list_blob_table("list_merge")
-    result = (
-        table.merge_insert("id")
-        .when_matched_update_all()
-        .when_not_matched_insert_all()
-        .execute([{"id": 1, "images": [b"one", b"two"]}, {"id": 2, "images": None}])
-    )
-    assert result.num_updated_rows == 1
-    assert result.num_inserted_rows == 1
-    hits = table.search().limit(10).to_arrow()
-    sizes = {
-        row["id"]: None if row["images"] is None else [d["size"] for d in row["images"]]
-        for row in hits.to_pylist()
-    }
-    assert sizes == {1: [3, 3], 2: None}
-
-
-def test_list_blob_column_queries_as_raw_descriptors():
-    table = _list_blob_table("list_query")
-    hits = table.search().limit(10).to_arrow()
-    element = hits.schema.field("images").type.value_type
-    assert pa.types.is_struct(element)
-    assert "_lance_row_id" not in element.names
-    with pytest.raises(ValueError, match="expected struct before segment"):
-        table.fetch_blobs("images.image", [0])
-
-
-def test_row_addressable_paths_exclude_list_children():
-    from lancedb.schema import row_addressable_blob_v2_paths
-
-    schema = pa.schema(
-        [
-            pa.field("id", pa.int64()),
-            pa.field("info", pa.struct([lancedb.blob("blob")])),
-            pa.field("images", pa.list_(lancedb.blob("image"))),
-        ]
-    )
-    assert blob_v2_column_paths(schema) == ["info.blob", "images.image"]
-    assert row_addressable_blob_v2_paths(schema) == ["info.blob"]
-
-
-def test_merge_insert_writes_pylance_blob_array():
-    table = _blob_table("merge_pylance", [{"id": 1, "image": b"before"}])
-    image = lance.blob_array([b"updated", b"inserted"])
-    assert type(image.type) is LanceBlobType
-    assert type(image.type) is type(lancedb.BlobType())
-    updates = pa.Table.from_arrays(
-        [pa.array([1, 2], type=pa.int64()), image], names=["id", "image"]
-    )
-
-    result = (
-        table.merge_insert("id")
-        .when_matched_update_all()
-        .when_not_matched_insert_all()
-        .execute(updates)
-    )
-
-    assert result.num_updated_rows == 1
-    assert result.num_inserted_rows == 1
-    by_id = _row_ids_by_id(table)
-    blobs = table.fetch_blobs("image", [by_id[1], by_id[2]])
-    assert blobs.to_pylist() == [b"updated", b"inserted"]
-
-
 def test_fetch_blobs_accepts_query_result():
    table = _blob_table("from_result", [{"id": 1, "image": b"gamma"}])
    hits = table.search().limit(10).to_arrow()
@@ -910,50 +403,6 @@ async def test_blob_v2_hybrid_fetch_blobs_async():
    assert {blobs[i].as_py() for i in range(len(blobs))} == {b"alpha", b"beta"}


-@pytest.mark.asyncio
-async def test_async_hybrid_typed_blob_projection_preserves_source_column():
-    db = await lancedb.connect_async("memory:///hybrid_typed_blob")
-    schema = pa.schema(
-        [
-            pa.field("id", pa.int64()),
-            pa.field("text", pa.utf8()),
-            pa.field("vector", pa.list_(pa.float32(), list_size=2)),
-            lancedb.blob("blob"),
-        ]
-    )
-    table = await db.create_table("hybrid_typed_blob", schema=schema)
-    await table.add(
-        [
-            {
-                "id": 1,
-                "text": "hello alpha",
-                "vector": [1.0, 0.0],
-                "blob": b"alpha",
-            },
-            {
-                "id": 2,
-                "text": "hello beta",
-                "vector": [0.9, 0.1],
-                "blob": b"beta",
-            },
-        ]
-    )
-    await table.create_index("text", config=FTS(with_position=False))
-
-    hits = await (
-        table.query()
-        .nearest_to([1.0, 0.0])
-        .nearest_to_text("hello")
-        .select({"blob_alias": col("blob")})
-        .limit(2)
-        .to_arrow()
-    )
-
-    assert "_lance_row_id" in hits.schema.field("blob_alias").type.names
-    blobs = await table.fetch_blobs("blob", hits)
-    assert {blobs[i].as_py() for i in range(len(blobs))} == {b"alpha", b"beta"}
-
-
 def test_blob_file_seek_read_and_read_range():
    payload = _identifiable_payload(1024)
    table = _blob_table("seek_read", [{"id": 1, "image": payload}])
@@ -1168,71 +617,3 @@ def test_fetch_blobs_nested_path_survives_sort_after_query():
 def _identifiable_payload(size: int) -> bytes:
    block = 256
    return b"".join(bytes([i % 256]) * block for i in range(size // block))
-
-
-def _external_uri_blob_array(uris):
-    blob_type = lancedb.blob("image").type
-    storage_type = blob_type.storage_type
-    child_names = [field.name for field in storage_type]
-    assert "uri" in child_names, "blob layout no longer has a uri child"
-    children = [
-        pa.array(uris if field.name == "uri" else [None] * len(uris), type=field.type)
-        for field in storage_type
-    ]
-    storage = pa.StructArray.from_arrays(children, fields=list(storage_type))
-    return pa.ExtensionArray.from_storage(blob_type, storage)
-
-
-def _external_uri_table_and_rows(name, uris):
-    db = lancedb.connect("memory:///")
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table(name, schema=schema)
-    rows = pa.Table.from_arrays(
-        [
-            pa.array(range(len(uris)), type=pa.int64()),
-            _external_uri_blob_array(uris),
-        ],
-        schema=schema,
-    )
-    return table, rows
-
-
-def test_add_external_uri_struct_round_trips_with_flag(tmp_path):
-    payload = b"external-uri-bytes"
-    blob_path = tmp_path / "payload.bin"
-    blob_path.write_bytes(payload)
-
-    table, rows = _external_uri_table_and_rows("external_struct", [blob_path.as_uri()])
-    table.add(rows, allow_external_blob_outside_bases=True)
-
-    hits = table.search().to_arrow()
-    blobs = table.fetch_blobs("image", hits)
-    assert blobs[0].as_py() == payload
-
-
-def test_add_external_uri_without_flag_raises(tmp_path):
-    blob_path = tmp_path / "payload.bin"
-    blob_path.write_bytes(b"unreachable")
-
-    table, rows = _external_uri_table_and_rows("external_no_flag", [blob_path.as_uri()])
-    with pytest.raises(ValueError, match="allow_external_blob_outside_bases"):
-        table.add(rows)
-    assert table.count_rows() == 0
-
-
-def test_add_external_uri_string_round_trips_with_flag(tmp_path):
-    payload = b"external-uri-bytes"
-    blob_path = tmp_path / "payload.bin"
-    blob_path.write_bytes(payload)
-
-    db = lancedb.connect("memory:///")
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    table = db.create_table("external_string", schema=schema)
-    table.add(
-        [{"id": 1, "image": blob_path.as_uri()}],
-        allow_external_blob_outside_bases=True,
-    )
-
-    hits = table.search().to_arrow()
-    blobs = table.fetch_blobs("image", hits)
-    assert blobs[0].as_py() == payload
@@ -52,7 +52,7 @@ class TestExprConstruction:
    def test_func(self):
        e = func("lower", col("name"))
        assert isinstance(e, Expr)
-        assert e.to_sql() == "lower(`name`)"
+        assert e.to_sql() == "lower(name)"

    def test_func_unknown_raises(self):
        with pytest.raises(Exception):
@@ -115,7 +115,7 @@ class TestExprOperators:
    def test_and_operator(self):
        e = (col("age") > lit(18)) & (col("status") == lit("active"))
        assert isinstance(e, Expr)
-        assert e.to_sql() == "((age > 18) AND (`status` = 'active'))"
+        assert e.to_sql() == "((age > 18) AND (status = 'active'))"

    def test_or_operator(self):
        e = (col("a") == lit(1)) | (col("b") == lit(2))
@@ -166,7 +166,7 @@ class TestExprOperators:
    def test_coerce_plain_str(self):
        e = col("name") == "alice"
        assert isinstance(e, Expr)
-        assert e.to_sql() == "(`name` = 'alice')"
+        assert e.to_sql() == "(name = 'alice')"

    def test_reflexive_comparisons(self):
        # 10 < col("age") swaps to col("age") > 10
@@ -198,85 +198,85 @@ class TestExprBytesLiteral:

    def test_bytes_equality_expr_sql(self):
        e = col("data") == lit(b"\xca\xfe")
-        assert e.to_sql() == "(`data` = X'CAFE')"
+        assert e.to_sql() == "(data = X'CAFE')"

    def test_bytes_ne_expr_sql(self):
        e = col("data") != lit(b"\xff")
-        assert e.to_sql() == "(`data` <> X'FF')"
+        assert e.to_sql() == "(data <> X'FF')"

    def test_bytes_compound_expr_sql(self):
        e = (col("data") == lit(b"\x01")) & (col("id") > lit(5))
-        assert e.to_sql() == "((`data` = X'01') AND (id > 5))"
+        assert e.to_sql() == "((data = X'01') AND (id > 5))"

    def test_bytes_in_function_call(self):
        # Regression test: binary literals inside scalar function calls
        # used to fail because DataFusion's unparser does not support Binary
        # scalars.  Now handled via a placeholder-substitution rewrite.
        e = func("contains", col("data"), lit(b"\xff"))
-        assert e.to_sql() == "contains(`data`, X'FF')"
+        assert e.to_sql() == "contains(data, X'FF')"

    def test_bytes_in_not(self):
        e = ~(col("data") == lit(b"\xff"))
-        assert e.to_sql() == "NOT (`data` = X'FF')"
+        assert e.to_sql() == "NOT (data = X'FF')"


 class TestExprStringMethods:
    def test_lower(self):
        e = col("name").lower()
        assert isinstance(e, Expr)
-        assert e.to_sql() == "lower(`name`)"
+        assert e.to_sql() == "lower(name)"

    def test_upper(self):
        e = col("name").upper()
        assert isinstance(e, Expr)
-        assert e.to_sql() == "upper(`name`)"
+        assert e.to_sql() == "upper(name)"

    def test_contains(self):
        e = col("text").contains(lit("hello"))
        assert isinstance(e, Expr)
-        assert e.to_sql() == "contains(`text`, 'hello')"
+        assert e.to_sql() == "contains(text, 'hello')"

    def test_contains_with_str_coerce(self):
        e = col("text").contains("hello")
        assert isinstance(e, Expr)
-        assert e.to_sql() == "contains(`text`, 'hello')"
+        assert e.to_sql() == "contains(text, 'hello')"

    def test_chained_lower_eq(self):
        e = col("name").lower() == lit("alice")
        assert isinstance(e, Expr)
-        assert e.to_sql() == "(lower(`name`) = 'alice')"
+        assert e.to_sql() == "(lower(name) = 'alice')"


 class TestExprCast:
    def test_cast_string(self):
        e = col("id").cast("string")
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(id, 'Utf8')"
+        assert e.to_sql() == "CAST(id AS VARCHAR)"

    def test_cast_int32(self):
        e = col("score").cast("int32")
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(score, 'Int32')"
+        assert e.to_sql() == "CAST(score AS INTEGER)"

    def test_cast_float64(self):
        e = col("val").cast("float64")
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(val, 'Float64')"
+        assert e.to_sql() == "CAST(val AS DOUBLE)"

    def test_cast_pyarrow_type(self):
        e = col("score").cast(pa.int32())
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(score, 'Int32')"
+        assert e.to_sql() == "CAST(score AS INTEGER)"

    def test_cast_pyarrow_float64(self):
        e = col("val").cast(pa.float64())
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(val, 'Float64')"
+        assert e.to_sql() == "CAST(val AS DOUBLE)"

    def test_cast_pyarrow_string(self):
        e = col("id").cast(pa.string())
        assert isinstance(e, Expr)
-        assert e.to_sql() == "arrow_cast(id, 'Utf8')"
+        assert e.to_sql() == "CAST(id AS VARCHAR)"

    def test_cast_pyarrow_and_string_equivalent(self):
        # pa.int32() and "int32" should produce equivalent SQL
@@ -597,14 +597,14 @@ class TestExprIsin:
    def test_isin_strs(self):
        assert (
            col("status").isin(["active", "pending"]).to_sql()
-            == "`status` IN ('active', 'pending')"
+            == "status IN ('active', 'pending')"
        )

    def test_isin_coerces_and_mixes(self):
        assert col("id").isin([lit(1), 2]).to_sql() == "id IN (1, 2)"

    def test_isin_empty(self):
-        assert col("id").isin([]).to_sql() == "false"
+        assert col("id").isin([]).to_sql() == "id IN ()"

    def test_isin_filter(self, simple_table):
        result = simple_table.search().where(col("id").isin([1, 3, 5])).to_arrow()
@@ -69,26 +69,6 @@ def _run_packaged(definition, *args):
    return namespace[definition.registration_request.artifact.entrypoint](*args)


-def test_udf_conda_environment():
-    @udf(conda=["scipy", "numpy"], conda_channels=["conda-forge", "defaults"])
-    def halve(value: float) -> float:
-        return value / 2
-
-    request = json.loads(halve.registration_request.to_canonical_json())
-    assert request["runtime"]["environment"] == {
-        "kind": "conda",
-        "packages": ["numpy", "scipy"],
-        "channels": ["conda-forge", "defaults"],
-    }
-    pip_request = json.loads(normalize_score.registration_request.to_canonical_json())
-    assert "channels" not in pip_request["runtime"]["environment"]
-
-    with pytest.raises(ValueError, match="not both"):
-        udf(name="both", pip=["numpy"], conda=["numpy"])(lambda value: value)
-    with pytest.raises(ValueError, match="requires conda"):
-        udf(name="channels", conda_channels=["conda-forge"])(lambda value: value)
-
-
 def test_udf_packages_attribute_access_and_body_imports():
    @udf
    def word_norm(body: str) -> float:
@@ -25,7 +25,6 @@ from lancedb.db import DBConnection
 from lancedb.index import FTS
 from lancedb.query import (
    BoostQuery,
-    DocumentGranularity,
    MatchQuery,
    MultiMatchQuery,
    PhraseQuery,
@@ -246,55 +245,6 @@ def test_create_inverted_index_rejects_invalid_block_size(table):
        table.create_index("text", config=FTS(block_size=129))


-def test_list_element_document_granularity(tmp_path):
-    docs_type = pa.list_(pa.struct([pa.field("content", pa.string())]))
-    docs = pa.array(
-        [
-            [
-                {"content": "alpha beta"},
-                None,
-                {"content": ""},
-                {"content": "the and"},
-                {"content": "alpha beta"},
-            ]
-        ],
-        type=docs_type,
-    )
-    table = ldb.connect(tmp_path).create_table(
-        "list_element_docs", pa.table({"id": [0], "docs": docs})
-    )
-    row_table = ldb.connect(tmp_path).create_table(
-        "row_docs", pa.table({"id": [0], "docs": docs})
-    )
-    row_table.create_index("docs.content", config=FTS())
-    row_result = row_table.search(MatchQuery("alpha", "docs.content")).to_arrow()
-    assert row_result.num_rows == 1
-    assert "_doc_index" not in row_result.column_names
-
-    granularity = DocumentGranularity.LIST_ELEMENT
-    table.create_index(
-        "docs.content",
-        config=FTS(with_position=True, document_granularity=granularity),
-    )
-    assert table.list_indices()[0].columns == ["docs.content"]
-
-    def coordinates(query):
-        result = table.search(query).limit(10).to_arrow()
-        doc_index_type = result.schema.field("_doc_index").type
-        assert pa.types.is_list(doc_index_type)
-        assert doc_index_type.value_type == pa.uint32()
-        return sorted(result["_doc_index"].to_pylist())
-
-    assert coordinates(
-        MatchQuery("alpha", "docs.content", document_granularity=granularity)
-    ) == [[0], [4]]
-    assert coordinates(
-        PhraseQuery("alpha beta", "docs.content", document_granularity=granularity)
-    ) == [[0], [4]]
-    assert coordinates(MatchQuery("alpha", "docs.content")) == [[0], [4]]
-    assert FTS().document_granularity is DocumentGranularity.ROW
-
-
 def test_create_inverted_index_respects_build_memory_limit(table):
    with pytest.raises(ValueError, match="exceeds worker memory limit"):
        table.create_index(
@@ -1139,20 +1089,6 @@ def test_fts_query_to_json():
    )
    assert json_str == expected

-    # Test MatchQuery with list-element document granularity
-    match_query = MatchQuery(
-        "hello world",
-        "text",
-        document_granularity=DocumentGranularity.LIST_ELEMENT,
-    )
-    json_str = match_query.to_json()
-    expected = (
-        '{"match":{"column":"text","terms":"hello world","boost":1.0,'
-        '"fuzziness":0,"max_expansions":50,"operator":"Or","prefix_length":0,'
-        '"document_granularity":"list_element"}}'
-    )
-    assert json_str == expected
-
    # Test MatchQuery with options
    match_query = MatchQuery("puppy", "text", fuzziness=2, boost=1.5, prefix_length=3)
    json_str = match_query.to_json()
@@ -1162,19 +1098,6 @@ def test_fts_query_to_json():
    )
    assert json_str == expected

-    # Test PhraseQuery with list-element document granularity
-    phrase_query = PhraseQuery(
-        "quick brown fox",
-        "title",
-        document_granularity=DocumentGranularity.LIST_ELEMENT,
-    )
-    json_str = phrase_query.to_json()
-    expected = (
-        '{"phrase":{"column":"title","terms":"quick brown fox","slop":0,'
-        '"document_granularity":"list_element"}}'
-    )
-    assert json_str == expected
-
    # Test PhraseQuery
    phrase_query = PhraseQuery("quick brown fox", "title")
    json_str = phrase_query.to_json()
@@ -675,21 +675,6 @@ def test_distance_range(table: lancedb.table.Table):
    assert res["_distance"].to_pylist() == [min_dist, max_dist]


-@pytest.mark.parametrize("expression", ["1 - _distance", "1.0 - _distance"])
-def test_select_arithmetic_with_distance(table, expression):
-    result = (
-        table.search([10, 10])
-        .select({"similarity": expression, "_distance": "_distance"})
-        .distance_type("cosine")
-        .to_arrow()
-    )
-
-    assert result.schema.field("similarity").type == pa.float32()
-    assert result["similarity"].to_pylist() == pytest.approx(
-        [1 - distance for distance in result["_distance"].to_pylist()]
-    )
-
-
@pytest.mark.asyncio
 async def test_distance_range_async(table_async: AsyncTable):
    q = [0, 0]
@@ -912,23 +897,6 @@ def test_query_builder_batches(table):
    assert rs_list["id"][1] == 2


-def test_batch_vector_query_shares_filtered_flat_scan(table):
-    query = (
-        table.search([[1.0, 2.0], [3.0, 4.0]])
-        .where("id > 0", prefilter=True)
-        .limit(1)
-        .select(["id"])
-    )
-
-    plan = query.explain_plan(verbose=True)
-    assert "KNNVectorDistance: queries=2" in plan
-    assert "UnionExec" not in plan
-
-    results = query.to_arrow()
-    assert len(results) == 2
-    assert results["query_index"].to_pylist() == [0, 1]
-
-
 def test_dynamic_projection(table):
    rs = (
        LanceVectorQueryBuilder(table, [0, 0], "vector")
@@ -1618,49 +1618,6 @@ def test_query_sync_fts():
        )


-def test_query_sync_fts_document_granularity():
-    from lancedb.query import DocumentGranularity, MatchQuery
-
-    def handler(body):
-        assert body == {
-            "full_text_query": {
-                "query": {
-                    "match": {
-                        "column": "docs.content",
-                        "terms": "alpha",
-                        "boost": 1.0,
-                        "fuzziness": 0,
-                        "max_expansions": 50,
-                        "operator": "Or",
-                        "prefix_length": 0,
-                        "document_granularity": "list_element",
-                    }
-                }
-            },
-            "k": 10,
-            "prefilter": True,
-            "vector": [],
-            "version": None,
-        }
-        return pa.table(
-            {
-                "id": [1, 1],
-                "_doc_index": pa.array([[0], [4]], type=pa.list_(pa.uint32())),
-            }
-        )
-
-    with query_test_table(handler, server_version=Version("0.6.0")) as table:
-        result = table.search(
-            MatchQuery(
-                "alpha",
-                "docs.content",
-                document_granularity=DocumentGranularity.LIST_ELEMENT,
-            )
-        ).to_arrow()
-
-    assert result["_doc_index"].to_pylist() == [[0], [4]]
-
-
 def test_query_sync_hybrid():
    def handler(body):
        if "full_text_query" in body:
@@ -4,7 +4,6 @@

 import asyncio
 import copy
-from concurrent.futures import ThreadPoolExecutor
 from datetime import timedelta
 import threading

@@ -87,25 +86,6 @@ def test_s3_lifecycle(s3_bucket: str):
    asyncio.run(test())


-@pytest.mark.s3_test
-def test_concurrent_open_table(s3_bucket: str):
-    uri = f"s3://{s3_bucket}/test_concurrent_open_table"
-    db = lancedb.connect(uri, storage_options=copy.copy(CONFIG))
-    db.create_table("test", pa.table({"x": [1, 2, 3]}))
-
-    num_workers = 32
-    barrier = threading.Barrier(num_workers)
-
-    def open_and_count(_):
-        barrier.wait()
-        return db.open_table("test").count_rows()
-
-    with ThreadPoolExecutor(max_workers=num_workers) as pool:
-        row_counts = list(pool.map(open_and_count, range(num_workers)))
-
-    assert row_counts == [3] * num_workers
-
-
@pytest.fixture()
 def kms_key():
    kms = get_boto3_client("kms", endpoint_url=CONFIG["aws_endpoint"])
@@ -11,9 +11,8 @@ import warnings
 import weakref
 from concurrent.futures import ThreadPoolExecutor
 from datetime import date, datetime, timedelta
-from decimal import Decimal
 from time import sleep
-from typing import Any, List
+from typing import List
 from unittest.mock import patch

 import lancedb
@@ -337,21 +336,6 @@ async def test_update_async(mem_db_async: AsyncConnection):
    assert await table.count_rows("id == 10") == 1


-@pytest.mark.asyncio
-async def test_update_expr_filter_literals_async(mem_db_async: AsyncConnection):
-    values = ["5", "4.66e-84", "it's"]
-    table = await mem_db_async.create_table(
-        "update_expr_literals",
-        data=[{"field": value, "result": "original"} for value in values],
-    )
-
-    for value in values:
-        update_res = await table.update({"result": value}, where=col("field") == value)
-        assert update_res.rows_updated == 1
-
-    assert (await table.to_arrow())["result"].to_pylist() == values
-
-
 def test_create_table(mem_db: DBConnection):
    schema = pa.schema(
        {
@@ -2359,148 +2343,6 @@ def test_update(mem_db: DBConnection):
    assert np.allclose(v, np.array([[1.2, 1.9], [1.1, 1.1]]))


-def test_update_expr_filter_literals(mem_db: DBConnection):
-    values = ["5", "4.66e-84", "it's"]
-    table = mem_db.create_table(
-        "update_expr_literals",
-        data=[{"field": value, "result": "original"} for value in values],
-    )
-
-    for value in values:
-        update_res = table.update(where=col("field") == value, values={"result": value})
-        assert update_res.rows_updated == 1
-
-    assert table.to_arrow()["result"].to_pylist() == values
-
-
-def test_update_expr_filter_preserves_typed_semantics(mem_db: DBConnection):
-    low = Decimal("1.234567890123456789")
-    high = Decimal("1.234567890123456790")
-    decimal_schema = pa.schema(
-        [("val", pa.decimal128(19, 18)), ("result", pa.string())]
-    )
-    decimal_table = mem_db.create_table(
-        "update_expr_decimal",
-        pa.table(
-            {"val": [low, high], "result": ["old", "old"]},
-            schema=decimal_schema,
-        ),
-    )
-    predicate = col("val") < lit(high)
-    assert decimal_table.search().where(predicate).to_arrow().num_rows == 1
-    result = decimal_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    keyword_table = mem_db.create_table(
-        "update_expr_keyword", [{"null": 1, "result": "old"}]
-    )
-    predicate = col("null") == 1
-    assert keyword_table.search().where(predicate).to_arrow().num_rows == 1
-    result = keyword_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    empty_in_table = mem_db.create_table(
-        "update_expr_empty_in", [{"id": 1, "result": "old"}]
-    )
-    predicate = col("id").isin([])
-    assert empty_in_table.search().where(predicate).to_arrow().num_rows == 0
-    result = empty_in_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 0
-
-    marker = "__lancedb_binary_placeholder_0__"
-    binary_schema = pa.schema(
-        [("payload", pa.binary()), ("text", pa.string()), ("result", pa.string())]
-    )
-    binary_table = mem_db.create_table(
-        "update_expr_binary",
-        pa.table(
-            {
-                "payload": [b"\x01", b"\x02"],
-                "text": ["other", marker],
-                "result": ["old", "old"],
-            },
-            schema=binary_schema,
-        ),
-    )
-    predicate = (col("payload") == lit(b"\x01")) | (col("text") == marker)
-    assert binary_table.search().where(predicate).to_arrow().num_rows == 2
-    result = binary_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 2
-
-    nonfinite_table = mem_db.create_table(
-        "update_expr_nonfinite",
-        [{"x": 1.0, "result": "old"}, {"x": 2.0, "result": "old"}],
-    )
-    predicate = col("x") < float("inf")
-    assert nonfinite_table.search().where(predicate).to_arrow().num_rows == 2
-    result = nonfinite_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 2
-
-    float16_table = mem_db.create_table(
-        "update_expr_float16",
-        [{"x": 1.0, "result": "old"}, {"x": 3.0, "result": "old"}],
-    )
-    predicate = col("x").cast(pa.float16()) < 2.0
-    assert float16_table.search().where(predicate).to_arrow().num_rows == 1
-    result = float16_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    string_cast_table = mem_db.create_table(
-        "update_expr_string_cast",
-        [{"x": 1, "result": "old"}, {"x": 2, "result": "old"}],
-    )
-    predicate = col("x").cast("string") == "1"
-    assert string_cast_table.search().where(predicate).to_arrow().num_rows == 1
-    result = string_cast_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    quoted_identifier_schema = pa.schema(
-        [("payload", pa.binary()), ("odd'name", pa.int64()), ("result", pa.string())]
-    )
-    quoted_identifier_table = mem_db.create_table(
-        "update_expr_quoted_identifier",
-        pa.table(
-            {"payload": [b"\x01"], "odd'name": [1], "result": ["old"]},
-            schema=quoted_identifier_schema,
-        ),
-    )
-    predicate = (col("payload") == lit(b"\x01")) & (col("odd'name") == 1)
-    assert quoted_identifier_table.search().where(predicate).to_arrow().num_rows == 1
-    result = quoted_identifier_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    decimal256_schema = pa.schema(
-        [("val", pa.decimal256(40, 2)), ("result", pa.string())]
-    )
-    decimal256_table = mem_db.create_table(
-        "update_expr_decimal256",
-        pa.table(
-            {
-                "val": [Decimal("1.00"), Decimal("3.00")],
-                "result": ["old", "old"],
-            },
-            schema=decimal256_schema,
-        ),
-    )
-    predicate = col("val") < lit(Decimal("2.00")).cast(pa.decimal256(40, 2))
-    assert decimal256_table.search().where(predicate).to_arrow().num_rows == 1
-    result = decimal256_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 1
-
-    binary_empty_table = mem_db.create_table(
-        "update_expr_binary_empty",
-        pa.table(
-            {"payload": [b"\x01", b"\x02"], "result": ["old", "old"]},
-            schema=pa.schema([("payload", pa.binary()), ("result", pa.string())]),
-        ),
-    )
-    predicate = (col("payload") == lit(b"\x01")).isin([])
-    assert binary_empty_table.search().where(predicate).to_arrow().num_rows == 0
-    assert predicate.to_sql() == "false"
-    result = binary_empty_table.update(where=predicate, values={"result": "new"})
-    assert result.rows_updated == 0
-
-
 def test_update_with_arrow_scalar(mem_db: DBConnection):
    schema = pa.schema({"id": pa.int64(), "vector": pa.list_(pa.float32(), 4)})
    table = mem_db.create_table("my_table", schema=schema)
@@ -3876,100 +3718,6 @@ async def test_optimize(mem_db_async: AsyncConnection):
    assert await table.query().to_arrow() == pa.table({"x": [[1], [2]]})


-@pytest.mark.asyncio
-async def test_optimize_compaction_options(mem_db_async: AsyncConnection):
-    table = await mem_db_async.create_table("test", data=[{"x": 1}])
-    await table.add([{"x": 2}])
-
-    stats = await table.optimize(
-        compaction_options={
-            "target_rows_per_fragment": 1,
-            "batch_size": 1,
-            "num_threads": 1,
-        }
-    )
-    assert stats.compaction.fragments_removed == 0
-    assert stats.compaction.fragments_added == 0
-
-    stats = await table.optimize(compaction_options={"target_rows_per_fragment": 3})
-    assert stats.compaction.fragments_removed == 2
-    assert stats.compaction.fragments_added == 1
-
-    with pytest.raises(ValueError, match="Invalid compaction option: unknown"):
-        await table.optimize(compaction_options={"unknown": 1})
-
-
-@pytest.mark.parametrize("option", ["max_source_rows", "max_source_bytes"])
-@pytest.mark.asyncio
-async def test_optimize_compaction_source_limits(
-    mem_db_async: AsyncConnection, option: str
-):
-    table = await mem_db_async.create_table("test", data=[{"x": 1}])
-    await table.add([{"x": 2}])
-
-    stats = await table.optimize(
-        compaction_options={
-            "target_rows_per_fragment": 3,
-            option: 1,
-        }
-    )
-    assert stats.compaction.fragments_removed == 0
-    assert stats.compaction.fragments_added == 0
-
-
-@pytest.mark.asyncio
-async def test_optimize_compaction_excluded_fragments(mem_db_async: AsyncConnection):
-    table = await mem_db_async.create_table("test", data=[{"x": 1}])
-    await table.add([{"x": 2}])
-
-    stats = await table.optimize(
-        compaction_options={
-            "target_rows_per_fragment": 3,
-            "excluded_fragment_ids": [0],
-        }
-    )
-    assert stats.compaction.fragments_removed == 0
-    assert stats.compaction.fragments_added == 0
-
-    stats = await table.optimize(compaction_options={"target_rows_per_fragment": 3})
-    assert stats.compaction.fragments_removed == 2
-    assert stats.compaction.fragments_added == 1
-
-
-@pytest.mark.parametrize(
-    ("option", "value", "message"),
-    [
-        ("target_rows_per_fragment", 0, "must be between 1 and 4294967295"),
-        ("max_rows_per_group", 0, "must be between 1 and 4294967295"),
-        ("batch_size", 0, "must be between 1 and 4294967295"),
-        ("num_threads", 0, "must be greater than 0"),
-        ("target_rows_per_fragment", 2**32, "must be between 1 and 4294967295"),
-        ("max_rows_per_group", 2**32, "must be between 1 and 4294967295"),
-        ("batch_size", 2**32, "must be between 1 and 4294967295"),
-        ("io_buffer_size", 2**63, "must be at most 9223372036854775807"),
-        ("max_source_rows", 0, "must be greater than 0"),
-        ("max_source_bytes", 0, "must be greater than 0"),
-        (
-            "excluded_fragment_ids",
-            [-1],
-            "must contain values between 0 and 4294967295",
-        ),
-        (
-            "excluded_fragment_ids",
-            [2**32],
-            "must contain values between 0 and 4294967295",
-        ),
-    ],
-)
-@pytest.mark.asyncio
-async def test_optimize_compaction_options_validation(
-    mem_db_async: AsyncConnection, option: str, value: Any, message: str
-):
-    table = await mem_db_async.create_table("test", data=[{"x": 1}])
-    with pytest.raises(ValueError, match=message):
-        await table.optimize(compaction_options={option: value})
-
-
@pytest.mark.asyncio
 async def test_optimize_delete_unverified(tmp_db_async: AsyncConnection, tmp_path):
    table = await tmp_db_async.create_table(
@@ -4181,29 +3929,6 @@ def test_computed_column_rejects_transforms_and_computed_together(tmp_path):
        table.add_columns({"a": "x + 1"}, computed={"b": "x * 2"})


-def test_computed_column_blob_projection_inherits_semantics(tmp_path):
-    schema = pa.schema([pa.field("id", pa.int64()), lancedb.blob("image")])
-    db = lancedb.connect(tmp_path)
-    table = db.create_table("computed_column_blob", schema=schema)
-    table.add(
-        [
-            {"id": 1, "image": b"hello"},
-            {"id": 2, "image": b""},
-            {"id": 3, "image": None},
-        ]
-    )
-
-    table.add_columns(computed={"image_copy": "image", "second_copy": "image_copy"})
-    assert table.refresh_column("image_copy").rows_filled == 2
-    assert table.refresh_column("second_copy").rows_filled == 2
-    assert table.blob_columns() == ["image", "image_copy", "second_copy"]
-
-    hits = table.search().with_row_id(True).limit(10).to_arrow()
-    rows = sorted(zip(hits["id"].to_pylist(), hits["_rowid"].to_pylist()))
-    copied = table.fetch_blobs("second_copy", [row_id for _, row_id in rows])
-    assert copied.to_pylist() == [b"hello", b"", None]
-
-
@pytest.mark.asyncio
 async def test_computed_column_async(tmp_path):
    db = await lancedb.connect_async(tmp_path)
@@ -7,7 +7,6 @@ import pathlib
 from typing import Optional

 import lance
-from lance.blob import BlobType as LanceBlobType
 from lancedb.conftest import MockTextEmbeddingFunction
 from lancedb.embeddings.base import EmbeddingFunctionConfig
 from lancedb.embeddings.registry import EmbeddingFunctionRegistry
@@ -908,165 +907,6 @@ def test_cast_to_target_schema():
    assert output == expected


-def test_cast_to_target_schema_coerces_binary_to_blob_v2():
-    data = pa.table({"image": pa.array([b"hello", None], type=pa.binary())})
-    target = pa.schema([lancedb.blob("image")])
-
-    output = _cast_to_target_schema(data.to_reader(), target).read_all()
-
-    image = output["image"].chunk(0)
-    assert type(image.type) is lancedb.BlobType
-    assert image.storage.to_pylist() == [
-        {"data": b"hello", "uri": None, "position": None, "size": None},
-        None,
-    ]
-
-
-def test_cast_to_target_schema_coerces_binary_to_metadata_blob_struct():
-    storage = lancedb.blob("image").type.storage_type
-    target = pa.schema(
-        [
-            pa.field(
-                "image",
-                storage,
-                metadata={
-                    b"ARROW:extension:name": b"lance.blob.v2",
-                    b"ARROW:extension:metadata": b"",
-                },
-            )
-        ]
-    )
-    data = pa.table({"image": pa.array([b"hello", None], type=pa.binary())})
-
-    output = _cast_to_target_schema(data.to_reader(), target).read_all()
-
-    image = output["image"].chunk(0)
-    assert not isinstance(image.type, pa.ExtensionType)
-    assert image.to_pylist() == [
-        {"data": b"hello", "uri": None, "position": None, "size": None},
-        None,
-    ]
-
-
-def test_cast_to_target_schema_coerces_nested_binary_blob():
-    data = pa.table(
-        {
-            "info": pa.array(
-                [{"blob": b"hello"}, {"blob": None}],
-                type=pa.struct([pa.field("blob", pa.binary())]),
-            )
-        }
-    )
-    target = pa.schema([pa.field("info", pa.struct([lancedb.blob("blob")]))])
-
-    output = _cast_to_target_schema(data.to_reader(), target).read_all()
-
-    blob = output["info"].chunk(0).field("blob")
-    assert type(blob.type) is lancedb.BlobType
-    assert blob.storage.to_pylist() == [
-        {"data": b"hello", "uri": None, "position": None, "size": None},
-        None,
-    ]
-
-
-def test_cast_to_target_schema_coerces_list_binary_blob_with_inferred_child_name():
-    data = pa.table(
-        {"images": pa.array([[b"a", b"b"], None], type=pa.list_(pa.binary()))}
-    )
-    target = pa.schema([pa.field("images", pa.list_(lancedb.blob("image")))])
-
-    output = _cast_to_target_schema(data.to_reader(), target).read_all()
-
-    images = output["images"].chunk(0)
-    assert images.type.value_field.name == "image"
-    assert type(images.type.value_type) is lancedb.BlobType
-    assert images.to_pylist()[1] is None
-    assert images.values.storage.to_pylist() == [
-        {"data": b"a", "uri": None, "position": None, "size": None},
-        {"data": b"b", "uri": None, "position": None, "size": None},
-    ]
-
-
-def test_list_blob_coercion_preserves_null_slots_with_nonzero_extent():
-    child = pa.field("image", pa.binary())
-    source = pa.ListArray.from_arrays(
-        pa.array([0, 2, 4], type=pa.int32()),
-        pa.array([b"a", b"b", b"dead", b"beef"], type=pa.binary()),
-        mask=pa.array([False, True]),
-    ).cast(pa.list_(child))
-    target = pa.schema([pa.field("images", pa.list_(lancedb.blob("image")))])
-
-    output = _cast_to_target_schema(
-        pa.table({"images": source}).to_reader(), target
-    ).read_all()
-
-    images = output["images"].chunk(0)
-    assert images.to_pylist()[1] is None
-    assert [b["data"] for b in images.to_pylist()[0]] == [b"a", b"b"]
-
-
-def test_fixed_size_list_blob_coercion_keeps_null_rows():
-    child = pa.field("frame", pa.binary())
-    source = (
-        pa.FixedSizeListArray.from_arrays(
-            pa.array([b"a", b"b", b"c", b"d"], type=pa.binary()), 2
-        )
-        .take(pa.array([0, None], type=pa.int32()))
-        .cast(pa.list_(child, 2))
-    )
-    target = pa.schema([pa.field("frames", pa.list_(lancedb.blob("frame"), 2))])
-
-    output = _cast_to_target_schema(
-        pa.table({"frames": source}).to_reader(), target
-    ).read_all()
-
-    frames = output["frames"].chunk(0)
-    assert frames.to_pylist()[1] is None
-    assert [b["data"] for b in frames.to_pylist()[0]] == [b"a", b"b"]
-
-
-def test_cast_to_target_schema_accepts_pylance_blob_v2():
-    target_type = lancedb.BlobType()
-    source = lance.blob_array([b"hello", None])
-    assert type(source.type) is LanceBlobType
-    assert type(source.type) is type(target_type)
-    data = pa.table({"image": source})
-    target = pa.schema([pa.field("image", target_type)])
-
-    output = _cast_to_target_schema(data.to_reader(), target).read_all()
-
-    image = output["image"].chunk(0)
-    assert type(image.type) is LanceBlobType
-    assert image.type == target_type
-    assert image.storage.to_pylist() == [
-        {"data": b"hello", "uri": None, "position": None, "size": None},
-        None,
-    ]
-
-
-def test_cast_to_target_schema_rejects_different_blob_v2_class():
-    class OtherBlobType(pa.ExtensionType):
-        def __init__(self):
-            super().__init__(lancedb.BlobType().storage_type, "lance.blob.v2")
-
-        def __arrow_ext_serialize__(self) -> bytes:
-            return b""
-
-        @classmethod
-        def __arrow_ext_deserialize__(
-            cls, storage_type: pa.DataType, serialized: bytes
-        ) -> "OtherBlobType":
-            return cls()
-
-    storage = lance.blob_array([b"hello"]).storage
-    source = pa.ExtensionArray.from_storage(OtherBlobType(), storage)
-    data = pa.table({"image": source})
-    target = pa.schema([lancedb.blob("image")])
-
-    with pytest.raises(pa.ArrowTypeError, match="different extension type"):
-        _cast_to_target_schema(data.to_reader(), target).read_all()
-
-
 def test_sanitize_data_stream():
    # Make sure we don't collect the whole stream when running sanitize_data
    schema = pa.schema({"a": pa.int32()})
@@ -130,14 +130,6 @@ impl PyExpr {

    // ── utilities ────────────────────────────────────────────────────────────

-    /// Return the referenced column name for a bare column expression.
-    fn column_name(&self) -> Option<String> {
-        match &self.0 {
-            DfExpr::Column(column) if column.relation.is_none() => Some(column.name.clone()),
-            _ => None,
-        }
-    }
-
    /// Render the expression as a SQL string (useful for debugging).
    fn to_sql(&self) -> PyResult<String> {
        lancedb::expr::expr_to_sql_string(&self.0).map_err(|e| PyValueError::new_err(e.to_string()))
@@ -8,7 +8,7 @@ use lancedb::index::vector::{
 };
 use lancedb::index::{
    Index as LanceDbIndex,
-    scalar::{BTreeIndexBuilder, DocumentGranularity, FmIndexBuilder, FtsIndexBuilder},
+    scalar::{BTreeIndexBuilder, FmIndexBuilder, FtsIndexBuilder},
 };
 use pyo3::IntoPyObject;
 use pyo3::types::PyStringMethods;
@@ -60,11 +60,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                    .ngram_min_length(params.ngram_min_length)
                    .ngram_max_length(params.ngram_max_length)
                    .ngram_prefix_only(params.prefix_only)
-                    .custom_stop_words(params.custom_stop_words)
-                    .document_granularity(
-                        DocumentGranularity::try_from(params.document_granularity.as_str())
-                            .map_err(|err| PyValueError::new_err(err.to_string()))?,
-                    );
+                    .custom_stop_words(params.custom_stop_words);
                if let Some(memory_limit) = params.memory_limit {
                    inner_opts = inner_opts.memory_limit_mb(memory_limit);
                }
@@ -225,7 +221,6 @@ struct FtsParams {
    block_size: usize,
    memory_limit: Option<u64>,
    num_workers: Option<usize>,
-    document_granularity: String,
 }

 #[derive(FromPyObject)]
@@ -486,7 +481,6 @@ mod tests {
    block_size = 128
    memory_limit = 2048
    num_workers = 7
-    document_granularity = 'row'

 config = FTS()",
                None,
@@ -1,7 +1,6 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors

-use std::collections::HashMap;
 use std::sync::Arc;
 use std::time::Duration;

@@ -17,8 +16,8 @@ use arrow::pyarrow::FromPyArrow;
 use arrow::pyarrow::IntoPyArrow;
 use arrow::pyarrow::ToPyArrow;
 use lancedb::index::scalar::{
-    BooleanQuery, BoostQuery, DocumentGranularity, FtsQuery, FullTextSearchQuery, MatchQuery,
-    MultiMatchQuery, Occur, Operator, PhraseQuery,
+    BooleanQuery, BoostQuery, FtsQuery, FullTextSearchQuery, MatchQuery, MultiMatchQuery, Occur,
+    Operator, PhraseQuery,
 };
 use lancedb::query::AnalyzePlanDistributedMetrics;
 use lancedb::query::QueryBase;
@@ -77,16 +76,8 @@ impl<'a, 'py> FromPyObject<'a, 'py> for PyLanceDB<FtsQuery> {
                let max_expansions = ob.getattr("max_expansions")?.extract()?;
                let operator = ob.getattr("operator")?.extract::<String>()?;
                let prefix_length = ob.getattr("prefix_length")?.extract()?;
-                let document_granularity = ob
-                    .getattr("document_granularity")?
-                    .extract::<Option<String>>()?
-                    .map(|value| {
-                        DocumentGranularity::try_from(value.as_str())
-                            .map_err(|err| PyValueError::new_err(err.to_string()))
-                    })
-                    .transpose()?;

-                let mut query =
+                Ok(Self(
                    MatchQuery::new(query)
                        .with_column(Some(column))
                        .with_boost(boost)
@@ -95,32 +86,21 @@ impl<'a, 'py> FromPyObject<'a, 'py> for PyLanceDB<FtsQuery> {
                        .with_operator(Operator::try_from(operator.as_str()).map_err(|e| {
                            PyValueError::new_err(format!("Invalid operator: {}", e))
                        })?)
-                        .with_prefix_length(prefix_length);
-                if let Some(document_granularity) = document_granularity {
-                    query = query.with_document_granularity(document_granularity);
-                }
-                Ok(Self(query.into()))
+                        .with_prefix_length(prefix_length)
+                        .into(),
+                ))
            }
            "PhraseQuery" => {
                let query = ob.getattr("query")?.extract()?;
                let column = ob.getattr("column")?.extract()?;
                let slop = ob.getattr("slop")?.extract()?;
-                let document_granularity = ob
-                    .getattr("document_granularity")?
-                    .extract::<Option<String>>()?
-                    .map(|value| {
-                        DocumentGranularity::try_from(value.as_str())
-                            .map_err(|err| PyValueError::new_err(err.to_string()))
-                    })
-                    .transpose()?;

-                let mut query = PhraseQuery::new(query)
-                    .with_column(Some(column))
-                    .with_slop(slop);
-                if let Some(document_granularity) = document_granularity {
-                    query = query.with_document_granularity(document_granularity);
-                }
-                Ok(Self(query.into()))
+                Ok(Self(
+                    PhraseQuery::new(query)
+                        .with_column(Some(column))
+                        .with_slop(slop)
+                        .into(),
+                ))
            }
            "BoostQuery" => {
                let positive: Self = ob.getattr("positive")?.extract()?;
@@ -187,13 +167,6 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
                kwargs.set_item("max_expansions", query.max_expansions)?;
                kwargs.set_item::<_, &str>("operator", query.operator.into())?;
                kwargs.set_item("prefix_length", query.prefix_length)?;
-                if let Some(document_granularity) = query.document_granularity {
-                    let value = match document_granularity {
-                        DocumentGranularity::Row => "row",
-                        DocumentGranularity::ListElement => "list_element",
-                    };
-                    kwargs.set_item("document_granularity", value)?;
-                }
                namespace
                    .getattr(intern!(py, "MatchQuery"))?
                    .call((query.terms, query.column.unwrap()), Some(&kwargs))
@@ -201,13 +174,6 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
            FtsQuery::Phrase(query) => {
                let kwargs = PyDict::new(py);
                kwargs.set_item("slop", query.slop)?;
-                if let Some(document_granularity) = query.document_granularity {
-                    let value = match document_granularity {
-                        DocumentGranularity::Row => "row",
-                        DocumentGranularity::ListElement => "list_element",
-                    };
-                    kwargs.set_item("document_granularity", value)?;
-                }
                namespace
                    .getattr(intern!(py, "PhraseQuery"))?
                    .call((query.terms, query.column.unwrap()), Some(&kwargs))
@@ -326,7 +292,6 @@ pub struct PyQueryRequest {
    pub filter: Option<PyQueryFilter>,
    pub full_text_search: Option<PyLanceDB<FtsQuery>>,
    pub select: PySelect,
-    pub select_source_columns: Option<HashMap<String, String>>,
    pub fast_search: Option<bool>,
    pub with_row_id: Option<bool>,
    pub use_lsm: Option<bool>,
@@ -357,7 +322,6 @@ impl From<AnyQuery> for PyQueryRequest {
                full_text_search: query_request
                    .full_text_search
                    .map(|fts| PyLanceDB(fts.query)),
-                select_source_columns: PySelect::source_columns(&query_request.select),
                select: PySelect(query_request.select),
                fast_search: Some(query_request.fast_search),
                with_row_id: Some(query_request.with_row_id),
@@ -383,7 +347,6 @@ impl From<AnyQuery> for PyQueryRequest {
                offset: vector_query.base.offset,
                filter: vector_query.base.filter.map(PyQueryFilter),
                full_text_search: None,
-                select_source_columns: PySelect::source_columns(&vector_query.base.select),
                select: PySelect(vector_query.base.select),
                fast_search: Some(vector_query.base.fast_search),
                with_row_id: Some(vector_query.base.with_row_id),
@@ -416,25 +379,6 @@ impl From<AnyQuery> for PyQueryRequest {
 #[derive(Clone)]
 pub struct PySelect(Select);

-impl PySelect {
-    fn source_columns(select: &Select) -> Option<HashMap<String, String>> {
-        match select {
-            Select::Expr(pairs) => Some(
-                pairs
-                    .iter()
-                    .filter_map(|(output, expr)| match expr {
-                        lancedb::expr::DfExpr::Column(column) if column.relation.is_none() => {
-                            Some((output.clone(), column.name.clone()))
-                        }
-                        _ => None,
-                    })
-                    .collect(),
-            ),
-            _ => None,
-        }
-    }
-}
-
 impl<'py> IntoPyObject<'py> for PySelect {
    type Target = PyAny;
    type Output = Bound<'py, Self::Target>;
@@ -20,9 +20,8 @@ use arrow::{
 use lancedb::blob::{BlobFile, BlobRangeRequest};
 use lancedb::index::scalar::FtsIndexBuilder;
 use lancedb::table::{
-    AddDataMode, ColumnAlteration, CompactionMode, CompactionOptions, Duration,
-    FieldMetadataUpdate, FtsToken as LanceDbFtsToken, IndexRemapMode, NewColumnTransform,
-    OptimizeAction, OptimizeOptions, Ref, Table as LanceDbTable,
+    AddDataMode, ColumnAlteration, Duration, FieldMetadataUpdate, FtsToken as LanceDbFtsToken,
+    NewColumnTransform, OptimizeAction, OptimizeOptions, Ref, Table as LanceDbTable,
 };
 use lancedb::tokenize as lancedb_tokenize;
 use pyo3::{
@@ -101,128 +100,6 @@ enum PredicateArg {
    Sql(String),
 }

-fn validate_positive_u32(value: u64, name: &str) -> PyResult<usize> {
-    if !(1..=u32::MAX as u64).contains(&value) {
-        return Err(PyValueError::new_err(format!(
-            "{name} must be between 1 and {}",
-            u32::MAX
-        )));
-    }
-    Ok(value as usize)
-}
-
-fn positive_u32(value: &Bound<'_, PyAny>, name: &str) -> PyResult<usize> {
-    validate_positive_u32(value.extract()?, name)
-}
-
-fn optional_positive_u32(value: &Bound<'_, PyAny>, name: &str) -> PyResult<Option<usize>> {
-    value
-        .extract::<Option<u64>>()?
-        .map(|value| validate_positive_u32(value, name))
-        .transpose()
-}
-
-fn optional_positive_usize(value: &Bound<'_, PyAny>, name: &str) -> PyResult<Option<usize>> {
-    let value: Option<usize> = value.extract()?;
-    if value == Some(0) {
-        return Err(PyValueError::new_err(format!(
-            "{name} must be greater than 0"
-        )));
-    }
-    Ok(value)
-}
-
-fn optional_positive_u64(value: &Bound<'_, PyAny>, name: &str) -> PyResult<Option<u64>> {
-    let value: Option<u64> = value.extract()?;
-    if value == Some(0) {
-        return Err(PyValueError::new_err(format!(
-            "{name} must be greater than 0"
-        )));
-    }
-    Ok(value)
-}
-
-fn u32_list(value: &Bound<'_, PyAny>, name: &str) -> PyResult<Vec<u32>> {
-    value
-        .extract::<Vec<i64>>()?
-        .into_iter()
-        .map(|value| {
-            u32::try_from(value).map_err(|_| {
-                PyValueError::new_err(format!(
-                    "{name} must contain values between 0 and {}",
-                    u32::MAX
-                ))
-            })
-        })
-        .collect()
-}
-
-fn optional_i64_bounded_u64(value: &Bound<'_, PyAny>, name: &str) -> PyResult<Option<u64>> {
-    let value: Option<u64> = value.extract()?;
-    if value.is_some_and(|value| value > i64::MAX as u64) {
-        return Err(PyValueError::new_err(format!(
-            "{name} must be at most {}",
-            i64::MAX
-        )));
-    }
-    Ok(value)
-}
-
-fn parse_compaction_options(options: Option<&Bound<'_, PyDict>>) -> PyResult<CompactionOptions> {
-    let mut parsed = CompactionOptions::default();
-    let Some(options) = options else {
-        return Ok(parsed);
-    };
-
-    for (key, value) in options.iter() {
-        let key: String = key.extract()?;
-        match key.as_str() {
-            "target_rows_per_fragment" => {
-                parsed.target_rows_per_fragment = positive_u32(&value, &key)?
-            }
-            "max_rows_per_group" => parsed.max_rows_per_group = positive_u32(&value, &key)?,
-            "max_bytes_per_file" => parsed.max_bytes_per_file = value.extract()?,
-            "materialize_deletions" => parsed.materialize_deletions = value.extract()?,
-            "materialize_deletions_threshold" => {
-                parsed.materialize_deletions_threshold = value.extract()?
-            }
-            "num_threads" => parsed.num_threads = optional_positive_usize(&value, &key)?,
-            "batch_size" => parsed.batch_size = optional_positive_u32(&value, &key)?,
-            "io_buffer_size" => parsed.io_buffer_size = optional_i64_bounded_u64(&value, &key)?,
-            "defer_index_remap" => parsed.defer_index_remap = value.extract()?,
-            "index_remap_mode" => {
-                let mode: String = value.extract()?;
-                parsed.index_remap_mode = IndexRemapMode::try_from(mode.as_str())
-                    .map_err(|err| PyValueError::new_err(err.to_string()))?;
-            }
-            "compaction_mode" => {
-                let mode: Option<String> = value.extract()?;
-                parsed.compaction_mode = mode
-                    .map(|mode| {
-                        CompactionMode::try_from(mode.as_str())
-                            .map_err(|err| PyValueError::new_err(err.to_string()))
-                    })
-                    .transpose()?;
-            }
-            "binary_copy_read_batch_bytes" => {
-                parsed.binary_copy_read_batch_bytes = value.extract()?
-            }
-            "max_source_fragments" => parsed.max_source_fragments = value.extract()?,
-            "max_source_rows" => parsed.max_source_rows = optional_positive_usize(&value, &key)?,
-            "max_source_bytes" => parsed.max_source_bytes = optional_positive_u64(&value, &key)?,
-            "excluded_fragment_ids" => parsed.excluded_fragment_ids = u32_list(&value, &key)?,
-            "max_overlays_per_fragment" => parsed.max_overlays_per_fragment = value.extract()?,
-            _ => {
-                return Err(PyValueError::new_err(format!(
-                    "Invalid compaction option: {key}"
-                )));
-            }
-        }
-    }
-
-    Ok(parsed)
-}
-
 /// Statistics about a compaction operation.
 #[pyclass(get_all, from_py_object)]
 #[derive(Clone, Debug)]
@@ -903,19 +780,15 @@ impl Table {
        })
    }

-    #[pyo3(signature = (data, mode, progress=None, write_parallelism=None, allow_external_blob_outside_bases=false))]
+    #[pyo3(signature = (data, mode, progress=None, write_parallelism=None))]
    pub fn add<'a>(
        self_: PyRef<'a, Self>,
        data: PyScannable,
        mode: String,
        progress: Option<Py<PyAny>>,
        write_parallelism: Option<usize>,
-        allow_external_blob_outside_bases: bool,
    ) -> PyResult<Bound<'a, PyAny>> {
-        let mut op = self_
-            .inner_ref()?
-            .add(data)
-            .allow_external_blob_outside_bases(allow_external_blob_outside_bases);
+        let mut op = self_.inner_ref()?.add(data);
        if mode == "append" {
            op = op.mode(AddDataMode::Append);
        } else if mode == "overwrite" {
@@ -1463,15 +1336,13 @@ impl Table {
    }

    /// Optimize the on-disk data by compacting and pruning old data, for better performance.
-    #[pyo3(signature = (cleanup_since_ms=None, delete_unverified=None, compaction_options=None))]
-    pub fn optimize<'py>(
-        self_: PyRef<'py, Self>,
+    #[pyo3(signature = (cleanup_since_ms=None, delete_unverified=None))]
+    pub fn optimize(
+        self_: PyRef<'_, Self>,
        cleanup_since_ms: Option<u64>,
        delete_unverified: Option<bool>,
-        compaction_options: Option<&Bound<'py, PyDict>>,
-    ) -> PyResult<Bound<'py, PyAny>> {
+    ) -> PyResult<Bound<'_, PyAny>> {
        let inner = self_.inner_ref()?.clone();
-        let compaction_options = parse_compaction_options(compaction_options)?;
        let older_than = if let Some(ms) = cleanup_since_ms {
            if ms > i64::MAX as u64 {
                return Err(PyValueError::new_err(format!(
@@ -1487,7 +1358,7 @@ impl Table {
        future_into_py(self_.py(), async move {
            let compaction_stats = inner
                .optimize(OptimizeAction::Compact {
-                    options: compaction_options,
+                    options: lancedb::table::CompactionOptions::default(),
                    remap_options: None,
                })
                .await
@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.38.0-beta.12"
+version = "0.38.0-beta.10"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true
@@ -13,7 +13,7 @@ use lance::dataset::{ReadParams, WriteMode, builder::DatasetBuilder};
 use lance::io::{ObjectStore, ObjectStoreParams, WrappingObjectStore};
 use lance_datafusion::utils::StreamingWriteSource;
 use lance_file::version::LanceFileVersion;
-use lance_io::object_store::{ReadDirOptions, StorageOptionsAccessor, StorageOptionsProvider};
+use lance_io::object_store::{StorageOptionsAccessor, StorageOptionsProvider};
 use lance_table::io::commit::commit_handler_from_url;
 use object_store::local::LocalFileSystem;
 use snafu::ResultExt;
@@ -281,22 +281,6 @@ impl std::fmt::Display for ListingDatabase {
 }

 const LANCE_EXTENSION: &str = "lance";
-
-/// The table a listed child of the database names, or `None` if the child is not a table.
-///
-/// A table is the directory `<name>.lance`; a loose file or any other directory under the
-/// database prefix belongs to something else. `dir_suffix` is `.lance`, built once by the
-/// caller rather than per child.
-/// The table a listed child directory holds, or `None` if it is not a table at all.
-///
-/// Only directories are considered, so a loose object named like a table is not one.
-fn table_name(location: &object_store::path::Path, dir_suffix: &str) -> Option<String> {
-    location
-        .filename()?
-        .strip_suffix(dir_suffix)
-        .map(String::from)
-        .filter(|name| !name.is_empty())
-}
 const ENGINE: &str = "engine";
 const MIRRORED_STORE: &str = "mirroredStore";

@@ -960,72 +944,51 @@ impl Database for ListingDatabase {
        Ok(f)
    }

-    /// List the tables in the database, a page at a time.
-    ///
-    /// The page_token is opaque, unlike the `start_after` parameter of [`Self::table_names()`].
-    ///
-    /// When there are no more results, the returned page_token will be None.
-    ///
-    /// `limit` is the maximum number of tables to return in the response. But it is possible
-    /// for the response to contain fewer than `limit` tables, even when there are more tables
-    /// to return. Clients should check the returned page_token to determine if there are
-    /// more results, rather than relying on the number of tables returned.
-    ///
-    /// The order that results are returned in not guaranteed to be stable across calls,
-    /// so clients should not rely on it.
    async fn list_tables(&self, request: ListTablesRequest) -> Result<ListTablesResponse> {
        if request.id.as_ref().map(|v| !v.is_empty()).unwrap_or(false) {
            return self.namespace_database().list_tables(request).await;
        }
-        let limit = request.limit.map(|limit| limit.max(0) as usize);
-        let dir_suffix = format!(".{LANCE_EXTENSION}");
-        let mut tables = Vec::new();
-        let mut page_token = request.page_token.filter(|token| !token.is_empty());
+        let mut f = self
+            .object_store
+            .read_dir(self.base_path.clone())
+            .await?
+            .iter()
+            .map(Path::new)
+            .filter(|path| {
+                let is_lance = path
+                    .extension()
+                    .and_then(|e| e.to_str())
+                    .map(|e| e == LANCE_EXTENSION);
+                is_lance.unwrap_or(false)
+            })
+            .filter_map(|p| p.file_stem().and_then(|s| s.to_str().map(String::from)))
+            .collect::<Vec<String>>();
+        f.sort();

-        // A page of nothing: the store rejects a limit of zero, and no table was handed over
-        // for a token to resume after.
-        if limit == Some(0) {
-            return Ok(ListTablesResponse {
-                context: None,
-                tables,
-                page_token: None,
-            });
+        // Handle pagination with page_token
+        if let Some(ref page_token) = request.page_token {
+            let index = f
+                .iter()
+                .position(|name| name.as_str() > page_token.as_str())
+                .unwrap_or(f.len());
+            f.drain(0..index);
        }

-        loop {
-            // Ask only for what the page still has room for, so a database holding more
-            // than one page costs one request per page rather than one per table.
-            let listing = self
-                .object_store
-                .read_dir_page(
-                    self.base_path.clone(),
-                    ReadDirOptions {
-                        page_token: page_token.take(),
-                        limit: limit.map(|limit| limit - tables.len()),
-                    },
-                )
-                .await?;
-            page_token = listing.page_token;
-            // Only child directories can be tables, and the store already separates them
-            // out, so the objects in the page are not looked at.
-            tables.extend(
-                listing
-                    .result
-                    .common_prefixes
-                    .iter()
-                    .filter_map(|location| table_name(location, &dir_suffix)),
-            );
-            // Children that are not tables leave the page short of the limit, so keep
-            // going until the page is full or the database runs out.
-            if page_token.is_none() || limit.is_none_or(|limit| tables.len() >= limit) {
-                break;
+        // Determine if there's a next page. The token is the last name of this page,
+        // not the first of the next one: the next page resumes strictly after the
+        // token, so naming the next page's first entry would skip it.
+        let next_page_token = match request.limit {
+            Some(limit) if f.len() > limit as usize => {
+                f.truncate(limit as usize);
+                f.last().cloned()
            }
-        }
+            _ => None,
+        };

        Ok(ListTablesResponse {
            context: None,
-            tables,
-            page_token,
+            tables: f,
+            page_token: next_page_token,
        })
    }

@@ -1513,7 +1476,7 @@ mod tests {
    use crate::table::{AnyQuery, WriteOptions};
    use arrow_array::{Int32Array, RecordBatch, StringArray};
    use arrow_schema::{DataType, Field, Schema, SchemaRef};
-    use futures::{TryStreamExt, future::try_join_all, stream::once};
+    use futures::{TryStreamExt, stream::once};
    use std::path::PathBuf;
    use std::sync::Arc;
    use std::time::Duration;
@@ -1521,182 +1484,6 @@ mod tests {
    use tokio::sync::Barrier;
    use tokio::time::timeout;

-    async fn create_tables(db: &ListingDatabase, names: &[&str]) {
-        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
-        for name in names {
-            db.create_table(CreateTableRequest {
-                name: name.to_string(),
-                namespace_path: vec![],
-                data: Box::new(RecordBatch::new_empty(schema.clone())) as Box<dyn Scannable>,
-                mode: CreateTableMode::Create,
-                write_options: Default::default(),
-                location: None,
-                namespace_client: None,
-            })
-            .await
-            .unwrap();
-        }
-    }
-
-    /// Every table in the database, taken `limit` at a time, which is how a caller walks a
-    /// listing: the token ends the walk, never a short page.
-    async fn walk(db: &ListingDatabase, limit: Option<i32>) -> Vec<String> {
-        let mut seen = Vec::new();
-        let mut page_token = None;
-        loop {
-            let page = db
-                .list_tables(ListTablesRequest {
-                    limit,
-                    page_token,
-                    ..Default::default()
-                })
-                .await
-                .unwrap();
-            seen.extend(page.tables);
-            page_token = page.page_token;
-            if page_token.is_none() {
-                return seen;
-            }
-            assert!(
-                seen.len() < 100,
-                "the walk is serving tables more than once"
-            );
-        }
-    }
-
-    /// Paging with the returned token has to visit every table exactly once, whatever the
-    /// page size, with nothing lost or repeated at a boundary.
-    #[rstest::rstest]
-    #[tokio::test]
-    async fn test_list_tables_pages_over_every_table_once(#[values(1, 2, 3, 5, 10)] limit: i32) {
-        let (_tempdir, db) = setup_database().await;
-        create_tables(&db, &["a", "b", "c", "d", "e"]).await;
-
-        assert_eq!(walk(&db, Some(limit)).await, vec!["a", "b", "c", "d", "e"]);
-    }
-
-    /// The token is opaque: it is whatever resumes the store the database sits on, not a
-    /// table name. Callers hand it back and nothing else.
-    ///
-    /// Nothing validates a token, so one invented by a caller is read as a position rather
-    /// than refused — which is why the token has to come back from a previous page.
-    #[tokio::test]
-    async fn test_the_page_token_is_not_a_table_name() {
-        let (_tempdir, db) = setup_database().await;
-        create_tables(&db, &["a", "b", "c"]).await;
-
-        let page = db
-            .list_tables(ListTablesRequest {
-                limit: Some(1),
-                ..Default::default()
-            })
-            .await
-            .unwrap();
-
-        assert_eq!(page.tables, vec!["a"]);
-        let token = page.page_token.expect("two tables are still to come");
-        assert_ne!(token, "a");
-
-        // Handing it back is the only thing a caller does with it, and it resumes.
-        let rest = db
-            .list_tables(ListTablesRequest {
-                page_token: Some(token),
-                ..Default::default()
-            })
-            .await
-            .unwrap();
-        assert_eq!(rest.tables, vec!["b", "c"]);
-    }
-
-    /// A limit the listing does not fill leaves no token behind, so a caller paging by token
-    /// stops without asking for an empty page.
-    #[tokio::test]
-    async fn test_a_listing_that_runs_out_has_no_token() {
-        let (_tempdir, db) = setup_database().await;
-        create_tables(&db, &["a", "b"]).await;
-
-        let page = db
-            .list_tables(ListTablesRequest {
-                limit: Some(10),
-                ..Default::default()
-            })
-            .await
-            .unwrap();
-
-        assert_eq!(page.tables, vec!["a", "b"]);
-        assert_eq!(page.page_token, None);
-    }
-
-    /// An empty page token means "from the start", which is how a client looping on a token
-    /// spells its first request.
-    #[tokio::test]
-    async fn test_an_empty_page_token_lists_from_the_start() {
-        let (_tempdir, db) = setup_database().await;
-        create_tables(&db, &["a", "b"]).await;
-
-        let page = db
-            .list_tables(ListTablesRequest {
-                page_token: Some(String::new()),
-                ..Default::default()
-            })
-            .await
-            .unwrap();
-
-        assert_eq!(page.tables, vec!["a", "b"]);
-    }
-
-    /// Listing follows the order the object store lists directories in, so a name that
-    /// extends another comes first: the `-` of `users-archive.lance` sorts below the `.` of
-    /// `users.lance`. Pagination pushes its cursor into the list request, so it cannot report
-    /// an order other than the one it resumes in.
-    #[tokio::test]
-    async fn test_listing_order_follows_the_store_not_the_table_name() {
-        let (_tempdir, db) = setup_database().await;
-        create_tables(&db, &["users", "users-archive", "users.old"]).await;
-
-        assert_eq!(
-            walk(&db, None).await,
-            vec!["users-archive", "users", "users.old"]
-        );
-        // And paging reports the same order, so a walk sees each table once.
-        assert_eq!(
-            walk(&db, Some(1)).await,
-            vec!["users-archive", "users", "users.old"]
-        );
-    }
-
-    /// Only directories named `<name>.lance` are tables; loose files and other directories
-    /// under the database prefix are not. A page spent on them is filled from the next one,
-    /// so a page holding only non-tables does not read as an empty database.
-    #[tokio::test]
-    async fn test_listing_ignores_non_table_children() {
-        let (tempdir, db) = setup_database().await;
-        create_tables(&db, &["real"]).await;
-        std::fs::write(tempdir.path().join("aaa-loose.lance"), b"not a table").unwrap();
-        create_dir_all(tempdir.path().join("aaa-scratch")).unwrap();
-
-        let page = db
-            .list_tables(ListTablesRequest {
-                limit: Some(1),
-                ..Default::default()
-            })
-            .await
-            .unwrap();
-
-        assert_eq!(page.tables, vec!["real"]);
-    }
-
-    #[tokio::test]
-    async fn listing_ignores_empty_table_name() {
-        let (tempdir, db) = setup_database().await;
-        create_dir_all(tempdir.path().join(".lance")).unwrap();
-        let page = db.list_tables(ListTablesRequest::default()).await.unwrap();
-        assert!(
-            page.tables.is_empty(),
-            "invalid empty table name was listed"
-        );
-    }
-
    async fn setup_database() -> (tempfile::TempDir, ListingDatabase) {
        let tempdir = tempdir().unwrap();
        let uri = tempdir.path().to_str().unwrap();
@@ -1827,59 +1614,6 @@ mod tests {
        );
    }

-    #[tokio::test]
-    async fn test_concurrent_open_table_reuses_connection_object_store() {
-        let tempdir = tempdir().unwrap();
-        let uri = tempdir.path().to_str().unwrap();
-        let session = Arc::new(lance::session::Session::default());
-        let request = ConnectRequest {
-            uri: uri.to_string(),
-            #[cfg(feature = "remote")]
-            client_config: Default::default(),
-            options: Default::default(),
-            namespace_client_properties: Default::default(),
-            manifest_enabled: false,
-            read_consistency_interval: None,
-            session: Some(session.clone()),
-        };
-        let db = ListingDatabase::connect_with_options(&request)
-            .await
-            .unwrap();
-        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
-        db.create_table(CreateTableRequest {
-            name: "test".to_string(),
-            namespace_path: vec![],
-            data: Box::new(RecordBatch::new_empty(schema)) as Box<dyn Scannable>,
-            mode: CreateTableMode::Create,
-            write_options: Default::default(),
-            location: None,
-            namespace_client: None,
-        })
-        .await
-        .unwrap();
-
-        let before = session.store_registry().stats();
-        let opened_tables = try_join_all((0..32).map(|_| {
-            db.open_table(OpenTableRequest {
-                name: "test".to_string(),
-                namespace_path: vec![],
-                index_cache_size: None,
-                lance_read_params: None,
-                location: None,
-                namespace_client: None,
-                managed_versioning: None,
-            })
-        }))
-        .await
-        .unwrap();
-        let after = session.store_registry().stats();
-
-        assert_eq!(opened_tables.len(), 32);
-        assert_eq!(after.misses, before.misses);
-        assert_eq!(after.active_stores, before.active_stores);
-        assert!(after.hits >= before.hits + 32);
-    }
-
    #[tokio::test]
    async fn test_listing_database_root_ops_do_not_create_manifest() {
        let tempdir = tempdir().unwrap();
@@ -19,7 +19,6 @@

 mod sql;

-pub(crate) use sql::canonicalize_sql_predicate;
 pub use sql::expr_to_sql_string;

 use std::sync::Arc;
@@ -157,7 +156,7 @@ mod tests {
        use datafusion_common::ScalarValue;
        let expr = col("data").eq(lit(ScalarValue::Binary(Some(vec![0xca, 0xfe]))));
        let sql = expr_to_sql_string(&expr).unwrap();
-        assert_eq!(sql, "(`data` = X'CAFE')");
+        assert_eq!(sql, "(data = X'CAFE')");
    }

    #[test]
@@ -167,7 +166,7 @@ mod tests {
        let int_expr = col("id").gt(lit(5i64));
        let combined = bin_expr.and(int_expr);
        let sql = expr_to_sql_string(&combined).unwrap();
-        assert_eq!(sql, "((`data` = X'01') AND (id > 5))");
+        assert_eq!(sql, "((data = X'01') AND (id > 5))");
    }

    #[test]
@@ -185,7 +184,7 @@ mod tests {
        // serialized correctly (regression test for placeholder rewrite path).
        let expr = contains(col("data"), lit(ScalarValue::Binary(Some(vec![0xff]))));
        let sql = expr_to_sql_string(&expr).unwrap();
-        assert_eq!(sql, "contains(`data`, X'FF')");
+        assert_eq!(sql, "contains(data, X'FF')");
    }

    #[test]
@@ -196,7 +195,7 @@ mod tests {
            .eq(lit(ScalarValue::Binary(Some(vec![0xab, 0xcd]))))
            .not();
        let sql = expr_to_sql_string(&expr).unwrap();
-        assert_eq!(sql, "NOT (`data` = X'ABCD')");
+        assert_eq!(sql, "NOT (data = X'ABCD')");
    }

    #[test]
@@ -206,122 +205,6 @@ mod tests {
        assert!(sql.contains("IN"), "expected IN in: {}", sql);
    }

-    #[test]
-    fn test_empty_is_in() {
-        let expr = is_in(col("id"), vec![]);
-        assert_eq!(expr_to_sql_string(&expr).unwrap(), "false");
-    }
-
-    #[test]
-    fn test_empty_is_in_discards_binary_children() {
-        use datafusion_common::ScalarValue;
-
-        let expr = is_in(
-            col("payload").eq(lit(ScalarValue::Binary(Some(vec![0x01])))),
-            vec![],
-        );
-        assert_eq!(expr_to_sql_string(&expr).unwrap(), "false");
-    }
-
-    #[test]
-    fn test_keyword_identifier() {
-        let expr = col("null").eq(lit(1i64));
-        assert_eq!(expr_to_sql_string(&expr).unwrap(), "(`null` = 1)");
-    }
-
-    #[test]
-    fn test_decimal_literal_preserves_type() {
-        use datafusion_common::ScalarValue;
-
-        let expr = col("val").lt(lit(ScalarValue::Decimal128(
-            Some(1_234_567_890_123_456_790),
-            19,
-            18,
-        )));
-        let sql = expr_to_sql_string(&expr).unwrap();
-        assert_eq!(
-            sql,
-            "(val < arrow_cast('1.234567890123456790', 'Decimal128(19, 18)'))"
-        );
-    }
-
-    #[test]
-    fn test_non_finite_float_literal_preserves_type() {
-        let expr = col("x").lt(lit(f64::INFINITY));
-        assert_eq!(
-            expr_to_sql_string(&expr).unwrap(),
-            "(x < arrow_cast('inf', 'Float64'))"
-        );
-    }
-
-    #[test]
-    fn test_cast_uses_arrow_type_name() {
-        let string = expr_cast(col("x"), DataType::Utf8);
-        assert_eq!(
-            expr_to_sql_string(&string).unwrap(),
-            "arrow_cast(x, 'Utf8')"
-        );
-
-        let int32 = expr_cast(col("x"), DataType::Int32);
-        assert_eq!(
-            expr_to_sql_string(&int32).unwrap(),
-            "arrow_cast(x, 'Int32')"
-        );
-
-        let expr = expr_cast(col("x"), DataType::Float16).lt(lit(2.0));
-        assert_eq!(
-            expr_to_sql_string(&expr).unwrap(),
-            "(arrow_cast(x, 'Float16') < 2.0)"
-        );
-
-        let decimal = expr_cast(lit("2.00"), DataType::Decimal256(40, 2));
-        assert_eq!(
-            expr_to_sql_string(&decimal).unwrap(),
-            "arrow_cast('2.00', 'Decimal256(40, 2)')"
-        );
-    }
-
-    #[test]
-    fn test_binary_placeholder_does_not_rewrite_user_string() {
-        use datafusion_common::ScalarValue;
-
-        let marker = "__lancedb_binary_placeholder_0__";
-        let expr = col("payload")
-            .eq(lit(ScalarValue::Binary(Some(vec![0x01]))))
-            .or(col("text").eq(lit(marker)));
-        assert_eq!(
-            expr_to_sql_string(&expr).unwrap(),
-            "((payload = X'01') OR (`text` = '__lancedb_binary_placeholder_0__'))"
-        );
-    }
-
-    #[test]
-    fn test_binary_binding_skips_quoted_identifiers() {
-        use datafusion_common::ScalarValue;
-
-        let expr = col("payload")
-            .eq(lit(ScalarValue::Binary(Some(vec![0x01]))))
-            .and(col("odd'name").eq(lit(1i64)))
-            .and(col("odd`'name").eq(lit(2i64)));
-        assert_eq!(
-            expr_to_sql_string(&expr).unwrap(),
-            "(((payload = X'01') AND (`odd'name` = 1)) AND (`odd``'name` = 2))"
-        );
-    }
-
-    #[test]
-    fn test_binary_placeholder_collision_search_is_linear() {
-        use datafusion_common::ScalarValue;
-
-        let collision_shaped = format!("__lancedb_binary_placeholder_0__{}", "_".repeat(64_000));
-        let expr = col("payload")
-            .eq(lit(ScalarValue::Binary(Some(vec![0x01]))))
-            .and(col("text").eq(lit(collision_shaped.clone())));
-        let sql = expr_to_sql_string(&expr).unwrap();
-        assert!(sql.contains("X'01'"));
-        assert!(sql.contains(&format!("'{collision_shaped}'")));
-    }
-
    #[test]
    fn test_multiple_binary_literals() {
        use datafusion_common::ScalarValue;
@@ -1,27 +1,10 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors

-use std::{
-    any::TypeId,
-    collections::{HashMap, HashSet},
-};
-
-use arrow_array::types::{
-    Decimal32Type, Decimal64Type, Decimal128Type, Decimal256Type, DecimalType,
-};
-use arrow_schema::DataType;
 use datafusion_common::ScalarValue;
 use datafusion_common::tree_node::{Transformed, TreeNode, TreeNodeRecursion};
 use datafusion_expr::Expr;
-use datafusion_functions::core::expr_fn::{
-    arrow_cast as datafusion_arrow_cast, arrow_try_cast as datafusion_arrow_try_cast,
-};
-use datafusion_sql::sqlparser::{
-    dialect::{Dialect as SqlParserDialect, GenericDialect},
-    keywords::ALL_KEYWORDS,
-    tokenizer::{Token, Tokenizer},
-};
-use datafusion_sql::unparser::{self, dialect::Dialect as UnparserDialect};
+use datafusion_sql::unparser::{self, dialect::Dialect};

 /// Unparser dialect that matches the quoting style expected by the Lance SQL
 /// parser.  Lance uses backtick (`` ` ``) as the only delimited-identifier
@@ -36,74 +19,17 @@ use datafusion_sql::unparser::{self, dialect::Dialect as UnparserDialect};
 ///   lower-case by the SQL parser, which would break case-sensitive schemas).
 struct LanceSqlDialect;

-impl UnparserDialect for LanceSqlDialect {
+impl Dialect for LanceSqlDialect {
    fn identifier_quote_style(&self, identifier: &str) -> Option<char> {
-        let identifier_upper = identifier.to_ascii_uppercase();
-        let needs_quote =
-            (identifier_upper != "ID" && ALL_KEYWORDS.contains(&identifier_upper.as_str()))
-                || identifier.chars().any(|c| c.is_ascii_uppercase())
-                || !identifier.chars().enumerate().all(|(i, c)| {
-                    c == '_' || c.is_ascii_alphabetic() || (i > 0 && c.is_ascii_digit())
-                });
+        let needs_quote = identifier.chars().any(|c| c.is_ascii_uppercase())
+            || !identifier
+                .chars()
+                .enumerate()
+                .all(|(i, c)| c == '_' || c.is_ascii_alphabetic() || (i > 0 && c.is_ascii_digit()));
        if needs_quote { Some('`') } else { None }
    }
 }

-/// Lance's tokenizer dialect with SQL-standard double-quoted identifiers added.
-///
-/// Keep this deliberately small: Lance's parser wraps `GenericDialect` and
-/// delegates only identifier recognition, leaving every other dialect option at
-/// its default. In particular, `/*! ... */` remains an ordinary block comment.
-#[derive(Debug, Default)]
-struct PredicateDialect(GenericDialect);
-
-impl SqlParserDialect for PredicateDialect {
-    fn dialect(&self) -> TypeId {
-        self.0.dialect()
-    }
-
-    fn is_identifier_start(&self, ch: char) -> bool {
-        self.0.is_identifier_start(ch)
-    }
-
-    fn is_identifier_part(&self, ch: char) -> bool {
-        self.0.is_identifier_part(ch)
-    }
-
-    fn is_delimited_identifier_start(&self, ch: char) -> bool {
-        ch == '"' || ch == '`'
-    }
-}
-
-/// Canonicalize a raw SQL predicate for Lance's parser.
-///
-/// Lance wraps [`GenericDialect`] for identifier recognition while retaining the
-/// default dialect behavior for every other lexical option. [`PredicateDialect`]
-/// mirrors that contract and additionally recognizes `"` as an identifier
-/// delimiter, allowing this function to rewrite only those identifier tokens.
-pub fn canonicalize_sql_predicate(predicate: &str) -> crate::Result<String> {
-    let dialect = PredicateDialect::default();
-    let tokens = Tokenizer::new(&dialect, predicate)
-        .with_unescape(false)
-        .tokenize()
-        .map_err(|err| crate::Error::InvalidInput {
-            message: format!("invalid SQL predicate: {err}"),
-        })?;
-
-    Ok(tokens
-        .into_iter()
-        .map(|token| match token {
-            Token::Word(word) if word.quote_style == Some('"') => {
-                // with_unescape(false) retains doubled double quotes. Decode
-                // those before escaping any backticks for Lance's delimiter.
-                let identifier = word.value.replace("\"\"", "\"").replace('`', "``");
-                format!("`{identifier}`")
-            }
-            other => other.to_string(),
-        })
-        .collect())
-}
-
 /// Prefix for placeholder strings inserted in place of binary literals.  Chosen
 /// to be extremely unlikely to occur in user data.
 const BINARY_PLACEHOLDER_PREFIX: &str = "__lancedb_binary_placeholder_";
@@ -113,128 +39,24 @@ fn bytes_to_hex_sql(bytes: &[u8]) -> String {
    format!("X'{hex}'")
 }

-fn string_literals(expr: &Expr) -> HashSet<String> {
-    let mut literals = HashSet::new();
+/// Returns true if *expr* contains a `Binary` or `LargeBinary` scalar literal
+/// anywhere in its subtree.  DataFusion's SQL unparser cannot serialize those
+/// variants, so we route such expressions through a placeholder-substitution
+/// path that emits SQL `X'...'` byte-string literals.
+fn has_binary_literal(expr: &Expr) -> bool {
+    let mut found = false;
    let _ = expr.apply(&mut |e: &Expr| {
-        if let Expr::Literal(
-            ScalarValue::Utf8(Some(value))
-            | ScalarValue::LargeUtf8(Some(value))
-            | ScalarValue::Utf8View(Some(value)),
-            _,
-        ) = e
-        {
-            literals.insert(value.clone());
-        }
-        Ok(TreeNodeRecursion::Continue)
-    });
-    literals
-}
-
-fn typed_string_literal(value: String, data_type: DataType) -> Expr {
-    datafusion_arrow_cast(
-        Expr::Literal(ScalarValue::Utf8(Some(value)), None),
-        Expr::Literal(ScalarValue::Utf8(Some(data_type.to_string())), None),
-    )
-}
-
-fn next_binary_placeholder(user_strings: &HashSet<String>, next_id: &mut usize) -> String {
-    loop {
-        let placeholder = format!("{BINARY_PLACEHOLDER_PREFIX}{}__", *next_id);
-        *next_id += 1;
-        if !user_strings.contains(&placeholder) {
-            return placeholder;
-        }
-    }
-}
-
-fn bind_binary_literals(
-    sql: &str,
-    mut bindings: HashMap<String, Vec<u8>>,
-) -> crate::Result<String> {
-    let bytes = sql.as_bytes();
-    let mut output = Vec::with_capacity(bytes.len());
-    let mut index = 0;
-
-    // Walk SQL string tokens once. Placeholders are plain, unescaped string
-    // literals, so this remains linear even when user strings are large or
-    // deliberately resemble the placeholder prefix.
-    while index < bytes.len() {
-        if bytes[index] == b'`' {
-            let identifier_start = index;
-            index += 1;
-            let mut identifier_end = None;
-            while index < bytes.len() {
-                if bytes[index] == b'`' {
-                    if index + 1 < bytes.len() && bytes[index + 1] == b'`' {
-                        index += 2;
-                    } else {
-                        index += 1;
-                        identifier_end = Some(index);
-                        break;
-                    }
-                } else {
-                    index += 1;
-                }
-            }
-
-            let Some(identifier_end) = identifier_end else {
-                return Err(crate::Error::InvalidInput {
-                    message: "unterminated identifier while binding binary literal".to_string(),
-                });
-            };
-            output.extend_from_slice(&bytes[identifier_start..identifier_end]);
-            continue;
-        }
-
-        if bytes[index] != b'\'' {
-            output.push(bytes[index]);
-            index += 1;
-            continue;
-        }
-
-        let literal_start = index;
-        index += 1;
-        let content_start = index;
-        let mut escaped = false;
-        let mut content_end = None;
-        while index < bytes.len() {
-            if bytes[index] == b'\'' {
-                if index + 1 < bytes.len() && bytes[index + 1] == b'\'' {
-                    escaped = true;
-                    index += 2;
-                } else {
-                    content_end = Some(index);
-                    index += 1;
-                    break;
-                }
-            } else {
-                index += 1;
-            }
-        }
-
-        let Some(content_end) = content_end else {
-            return Err(crate::Error::InvalidInput {
-                message: "unterminated string while binding binary literal".to_string(),
-            });
-        };
-
-        let placeholder = &sql[content_start..content_end];
-        if !escaped && let Some(value) = bindings.remove(placeholder) {
-            output.extend_from_slice(bytes_to_hex_sql(&value).as_bytes());
+        if matches!(
+            e,
+            Expr::Literal(ScalarValue::Binary(_) | ScalarValue::LargeBinary(_), _)
+        ) {
+            found = true;
+            Ok(TreeNodeRecursion::Stop)
        } else {
-            output.extend_from_slice(&bytes[literal_start..index]);
+            Ok(TreeNodeRecursion::Continue)
        }
-    }
-
-    if !bindings.is_empty() {
-        return Err(crate::Error::InvalidInput {
-            message: "failed to bind binary literal while serializing expression".to_string(),
-        });
-    }
-
-    String::from_utf8(output).map_err(|e| crate::Error::InvalidInput {
-        message: format!("failed to bind binary literal: {e}"),
-    })
+    });
+    found
 }

 fn run_unparser(expr: &Expr) -> crate::Result<String> {
@@ -247,37 +69,25 @@ fn run_unparser(expr: &Expr) -> crate::Result<String> {
 }

 pub fn expr_to_sql_string(expr: &Expr) -> crate::Result<String> {
-    // DataFusion's unparser needs a few adaptations before its SQL can be
-    // reparsed by Lance without changing the typed expression's semantics:
-    //
-    // * decimal literals need an explicit cast to preserve precision and scale;
-    // * casts need exact Arrow type names rather than SQL type aliases;
-    // * an empty IN list is valid in DataFusion but invalid SQL;
-    // * binary literals are unsupported by the unparser and need placeholders.
-    // Eliminate empty membership expressions before visiting their children.
-    // Otherwise a discarded binary child could leave behind a stale binding.
+    // Fast path: no binary literals — DataFusion's unparser handles everything.
+    if !has_binary_literal(expr) {
+        return run_unparser(expr);
+    }
+
+    // Slow path: DataFusion's unparser cannot serialize `Binary`/`LargeBinary`
+    // scalars, so we rewrite each one to a unique string-literal placeholder,
+    // let the unparser do the rest of the work, then substitute the SQL
+    // `X'...'` byte-string literal back in.  This keeps the operator/function
+    // serialization logic centralized in DataFusion and works for every
+    // expression node type the unparser supports.
+    let mut bindings: Vec<Vec<u8>> = Vec::new();
    let rewritten = expr
        .clone()
-        .transform(|e: Expr| match e {
-            Expr::InList(in_list) if in_list.list.is_empty() => Ok(Transformed::yes(
-                Expr::Literal(ScalarValue::Boolean(Some(in_list.negated)), None),
-            )),
-            other => Ok(Transformed::no(other)),
-        })
-        .map_err(|e| crate::Error::InvalidInput {
-            message: format!("failed to rewrite expression: {e}"),
-        })?
-        .data;
-
-    let user_strings = string_literals(&rewritten);
-    let mut next_placeholder_id = 0;
-    let mut binary_bindings = HashMap::new();
-    let rewritten = rewritten
        .transform(|e: Expr| match e {
            Expr::Literal(ScalarValue::Binary(Some(bytes)), m)
            | Expr::Literal(ScalarValue::LargeBinary(Some(bytes)), m) => {
-                let placeholder = next_binary_placeholder(&user_strings, &mut next_placeholder_id);
-                binary_bindings.insert(placeholder.clone(), bytes);
+                let placeholder = format!("{}{}__", BINARY_PLACEHOLDER_PREFIX, bindings.len());
+                bindings.push(bytes);
                Ok(Transformed::yes(Expr::Literal(
                    ScalarValue::Utf8(Some(placeholder)),
                    m,
@@ -287,57 +97,6 @@ pub fn expr_to_sql_string(expr: &Expr) -> crate::Result<String> {
            | Expr::Literal(ScalarValue::LargeBinary(None), m) => {
                Ok(Transformed::yes(Expr::Literal(ScalarValue::Null, m)))
            }
-            Expr::Literal(ScalarValue::Decimal32(Some(value), precision, scale), _m) => {
-                let value = Decimal32Type::format_decimal(value, precision, scale);
-                Ok(Transformed::yes(typed_string_literal(
-                    value,
-                    DataType::Decimal32(precision, scale),
-                )))
-            }
-            Expr::Literal(ScalarValue::Decimal64(Some(value), precision, scale), _m) => {
-                let value = Decimal64Type::format_decimal(value, precision, scale);
-                Ok(Transformed::yes(typed_string_literal(
-                    value,
-                    DataType::Decimal64(precision, scale),
-                )))
-            }
-            Expr::Literal(ScalarValue::Decimal128(Some(value), precision, scale), _m) => {
-                let value = Decimal128Type::format_decimal(value, precision, scale);
-                Ok(Transformed::yes(typed_string_literal(
-                    value,
-                    DataType::Decimal128(precision, scale),
-                )))
-            }
-            Expr::Literal(ScalarValue::Decimal256(Some(value), precision, scale), _m) => {
-                let value = Decimal256Type::format_decimal(value, precision, scale);
-                Ok(Transformed::yes(typed_string_literal(
-                    value,
-                    DataType::Decimal256(precision, scale),
-                )))
-            }
-            Expr::Literal(ScalarValue::Float16(Some(value)), _m) if !value.is_finite() => Ok(
-                Transformed::yes(typed_string_literal(value.to_string(), DataType::Float16)),
-            ),
-            Expr::Literal(ScalarValue::Float32(Some(value)), _m) if !value.is_finite() => Ok(
-                Transformed::yes(typed_string_literal(value.to_string(), DataType::Float32)),
-            ),
-            Expr::Literal(ScalarValue::Float64(Some(value)), _m) if !value.is_finite() => Ok(
-                Transformed::yes(typed_string_literal(value.to_string(), DataType::Float64)),
-            ),
-            Expr::Cast(cast) => Ok(Transformed::yes(datafusion_arrow_cast(
-                *cast.expr,
-                Expr::Literal(
-                    ScalarValue::Utf8(Some(cast.field.data_type().to_string())),
-                    None,
-                ),
-            ))),
-            Expr::TryCast(cast) => Ok(Transformed::yes(datafusion_arrow_try_cast(
-                *cast.expr,
-                Expr::Literal(
-                    ScalarValue::Utf8(Some(cast.field.data_type().to_string())),
-                    None,
-                ),
-            ))),
            other => Ok(Transformed::no(other)),
        })
        .map_err(|e| crate::Error::InvalidInput {
@@ -345,58 +104,12 @@ pub fn expr_to_sql_string(expr: &Expr) -> crate::Result<String> {
        })?
        .data;

-    let sql = run_unparser(&rewritten)?;
-    if binary_bindings.is_empty() {
-        Ok(sql)
-    } else {
-        bind_binary_literals(&sql, binary_bindings)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::canonicalize_sql_predicate;
-
-    #[test]
-    fn normalizes_double_quoted_identifiers() {
-        assert_eq!(
-            canonicalize_sql_predicate(r#""PartyAbbrev" = 'D'"#).unwrap(),
-            "`PartyAbbrev` = 'D'"
-        );
-        assert_eq!(
-            canonicalize_sql_predicate(r#""MetaData"."userId" = 5"#).unwrap(),
-            "`MetaData`.`userId` = 5"
-        );
-        assert_eq!(
-            canonicalize_sql_predicate(r#""a""b" = 1"#).unwrap(),
-            "`a\"b` = 1"
-        );
-    }
-
-    #[test]
-    fn preserves_quotes_inside_literals_and_backticks() {
-        let filter = r#"name = 'Alice "Ace"' AND `quoted"field` = 1"#;
-        assert_eq!(canonicalize_sql_predicate(filter).unwrap(), filter);
-    }
-
-    #[test]
-    fn preserves_literals_and_comments_using_lance_dialect_rules() {
-        let predicate = r#"path = '\' AND "PartyAbbrev" = 'D' -- unmatched " in comment"#;
-        assert_eq!(
-            canonicalize_sql_predicate(predicate).unwrap(),
-            r#"path = '\' AND `PartyAbbrev` = 'D' -- unmatched " in comment"#
-        );
-
-        let predicate = r#"id = 1 /* unmatched " in block comment */"#;
-        assert_eq!(canonicalize_sql_predicate(predicate).unwrap(), predicate);
-
-        let predicate = r#"id = 1 /*! OR "PartyAbbrev" = 'D' */"#;
-        assert_eq!(canonicalize_sql_predicate(predicate).unwrap(), predicate);
-    }
-
-    #[test]
-    fn rejects_unterminated_double_quoted_identifier() {
-        let error = canonicalize_sql_predicate(r#""PartyAbbrev = 'D'"#).unwrap_err();
-        assert!(matches!(error, crate::Error::InvalidInput { .. }));
+    let mut sql = run_unparser(&rewritten)?;
+    for (i, bytes) in bindings.iter().enumerate() {
+        // The unparser quotes string literals with single quotes, so the
+        // placeholder appears as `'__lancedb_binary_placeholder_<i>__'`.
+        let quoted = format!("'{}{}__'", BINARY_PLACEHOLDER_PREFIX, i);
+        sql = sql.replace(&quoted, &bytes_to_hex_sql(bytes));
    }
+    Ok(sql)
 }
@@ -186,9 +186,6 @@ pub struct PythonEnvironmentSpec {
    pub kind: String,
    #[serde(default, skip_serializing_if = "Vec::is_empty")]
    pub packages: Vec<String>,
-    /// Conda channels in priority order; conda environments only.
-    #[serde(default, skip_serializing_if = "Vec::is_empty")]
-    pub channels: Vec<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub path: Option<String>,
    #[serde(default, skip_serializing_if = "Vec::is_empty")]
@@ -586,26 +583,3 @@ impl RefreshColumnResult {
 }

 impl_json!(RefreshColumnResult);
-
-#[cfg(test)]
-mod conda_environment_tests {
-    use super::PythonEnvironmentSpec;
-
-    #[test]
-    fn conda_channels_round_trip_and_pip_stays_bare() {
-        let conda: PythonEnvironmentSpec = serde_json::from_str(
-            r#"{"kind":"conda","packages":["numpy"],"channels":["conda-forge"]}"#,
-        )
-        .unwrap();
-        assert_eq!(conda.channels, ["conda-forge"]);
-        assert!(
-            serde_json::to_string(&conda)
-                .unwrap()
-                .contains(r#""channels":["conda-forge"]"#)
-        );
-
-        let pip: PythonEnvironmentSpec =
-            serde_json::from_str(r#"{"kind":"pip","packages":["numpy"]}"#).unwrap();
-        assert!(!serde_json::to_string(&pip).unwrap().contains("channels"));
-    }
-}
@@ -63,5 +63,4 @@ pub struct FmIndexBuilder {}
 pub use lance_index::scalar::FullTextSearchQuery;
 pub use lance_index::scalar::InvertedIndexParams as FtsIndexBuilder;
 pub use lance_index::scalar::InvertedIndexParams;
-pub use lance_index::scalar::inverted::DocumentGranularity;
 pub use lance_index::scalar::inverted::query::*;
@@ -10,7 +10,7 @@ use lance::io::WrappingObjectStore;
 use object_store::{
    CopyOptions, Error, GetOptions, GetResult, ListResult, MultipartUpload, ObjectMeta,
    ObjectStore, ObjectStoreExt, PutMultipartOptions, PutOptions, PutPayload, PutResult, Result,
-    UploadPart, list::PaginatedListStore, path::Path,
+    UploadPart, path::Path,
 };

 use async_trait::async_trait;
@@ -187,14 +187,6 @@ impl WrappingObjectStore for MirroringObjectStoreWrapper {
            secondary: self.secondary.clone(),
        })
    }
-
-    fn wrap_paginated(
-        &self,
-        _store_prefix: &str,
-        original: Arc<dyn PaginatedListStore>,
-    ) -> Option<Arc<dyn PaginatedListStore>> {
-        Some(original)
-    }
 }

 // windows pathing can't be simply concatenated
@@ -12,7 +12,7 @@ use lance::io::WrappingObjectStore;
 use object_store::{
    CopyOptions, GetOptions, GetResult, ListResult, MultipartUpload, ObjectMeta, ObjectStore,
    PutMultipartOptions, PutOptions, PutPayload, PutResult, RenameOptions, Result as OSResult,
-    UploadPart, list::PaginatedListStore, path::Path,
+    UploadPart, path::Path,
 };

 #[derive(Debug, Default)]
@@ -57,14 +57,6 @@ impl WrappingObjectStore for IoStatsHolder {
            stats: self.0.clone(),
        })
    }
-
-    fn wrap_paginated(
-        &self,
-        _store_prefix: &str,
-        original: Arc<dyn PaginatedListStore>,
-    ) -> Option<Arc<dyn PaginatedListStore>> {
-        Some(original)
-    }
 }

 impl IoTrackingStore {
@@ -47,10 +47,6 @@ impl TerminalResult {
        }
    }

-    pub(crate) fn value(&self) -> Option<&Value> {
-        self.value.as_ref()
-    }
-
    fn decode<T: DeserializeOwned>(self) -> Result<T> {
        let value = self.value.ok_or_else(|| match &self.request_id {
            Some(request_id) => Error::Http {
@@ -170,15 +170,6 @@ pub(crate) fn plan(
    filter: Option<&str>,
    limit: Option<u64>,
 ) -> Result<(MaterializedViewDefinition, Vec<ArrowField>, Lineage)> {
-    let filter = filter
-        .map(crate::expr::canonicalize_sql_predicate)
-        .transpose()
-        .map_err(|err| match err {
-            Error::InvalidInput { message } => Error::InvalidInput {
-                message: format!("invalid view filter: {message}"),
-            },
-            err => err,
-        })?;
    let projections: Vec<(String, String)> = if projections.is_empty() {
        source_schema
            .fields()
@@ -283,7 +274,7 @@ pub(crate) fn plan(
        declared.push(output);
    }

-    if let Some(filter) = filter.as_deref() {
+    if let Some(filter) = filter {
        let expr = planner
            .parse_filter(filter)
            .map_err(|e| Error::InvalidInput {
@@ -323,7 +314,7 @@ pub(crate) fn plan(
            .into_iter()
            .map(|(output, expression)| ViewProjection { output, expression })
            .collect(),
-        filter,
+        filter: filter.map(String::from),
        limit,
        inputs,
    };
@@ -46,9 +46,8 @@ use lance_table::format::Fragment;
 use serde::{Deserialize, Serialize};

 use super::{
-    DEFINITION_META_KEY, INCARNATION_META_KEY, MaterializedViewDefinition,
-    REFRESHED_AT_MS_META_KEY, SOURCE_ROW_ID_COLUMN, SOURCE_VERSION_META_KEY,
-    definition_to_metadata,
+    INCARNATION_META_KEY, MaterializedViewDefinition, REFRESHED_AT_MS_META_KEY,
+    SOURCE_ROW_ID_COLUMN, SOURCE_VERSION_META_KEY,
 };
 use crate::database::OpenTableRequest;
 use crate::table::{NativeTable, NativeTableExt, Table};
@@ -198,28 +197,8 @@ pub(crate) async fn execute_refresh(
            ),
        });
    }
-    let definition_changed =
-        definition.filter != replanned.filter || definition.inputs != replanned.inputs;
    let definition = &replanned;

-    // A watermark written for a legacy raw filter certifies the rows that
-    // filter produced, not the canonical predicate above. Rebuild instead of
-    // accepting or advancing it, and persist the migrated definition in the
-    // same metadata commit that certifies the replacement rows.
-    if definition_changed {
-        return rebuild(
-            view_native,
-            &view_ds,
-            &source_ds,
-            source_version,
-            source_ts,
-            definition,
-            true,
-            expected_incarnation,
-        )
-        .await;
-    }
-
    let metadata = &view_ds.schema().metadata;
    let watermark: Option<u64> = metadata
        .get(SOURCE_VERSION_META_KEY)
@@ -278,7 +257,6 @@ pub(crate) async fn execute_refresh(
                        source_version,
                        source_ts,
                        definition,
-                        false,
                        expected_incarnation,
                    )
                    .await
@@ -293,7 +271,6 @@ pub(crate) async fn execute_refresh(
                source_version,
                source_ts,
                definition,
-                false,
                expected_incarnation,
            )
            .await
@@ -706,7 +683,6 @@ async fn incremental(
            view_ds.clone(),
            source_version,
            source_ts,
-            None,
            expected_incarnation,
        )
        .await?;
@@ -728,7 +704,6 @@ async fn incremental(
            published,
            source_version,
            source_ts,
-            None,
            expected_incarnation,
        )
        .await?;
@@ -800,7 +775,6 @@ async fn incremental(
            published,
            source_version,
            source_ts,
-            None,
            expected_incarnation,
        )
        .await?;
@@ -850,14 +824,12 @@ async fn incremental(
        appended,
        source_version,
        source_ts,
-        None,
        expected_incarnation,
    )
    .await?;
    Ok(Some(result))
 }

-#[allow(clippy::too_many_arguments)]
 async fn rebuild(
    view_native: &NativeTable,
    view_ds: &Dataset,
@@ -865,7 +837,6 @@ async fn rebuild(
    source_version: u64,
    source_ts: u128,
    definition: &MaterializedViewDefinition,
-    persist_definition: bool,
    expected_incarnation: Option<&str>,
 ) -> Result<RefreshMaterializedViewResult> {
    let rows_written = Arc::new(AtomicU64::new(0));
@@ -896,7 +867,6 @@ async fn rebuild(
        replaced,
        source_version,
        source_ts,
-        persist_definition.then_some(definition),
        expected_incarnation,
    )
    .await?;
@@ -1011,7 +981,6 @@ async fn stamp_watermark(
    mut dataset: Dataset,
    source_version: u64,
    source_ts: u128,
-    definition: Option<&MaterializedViewDefinition>,
    expected_incarnation: Option<&str>,
 ) -> Result<u64> {
    ensure_incarnation(&dataset, expected_incarnation, dataset.uri()).await?;
@@ -1024,32 +993,27 @@ async fn stamp_watermark(
        .get(INCARNATION_META_KEY)
        .cloned()
        .unwrap_or_else(|| uuid::Uuid::new_v4().to_string());
-    let mut metadata = vec![(INCARNATION_META_KEY.to_string(), Some(incarnation))];
-    if let Some(definition) = definition {
-        metadata.push((
-            DEFINITION_META_KEY.to_string(),
-            Some(definition_to_metadata(definition)?),
-        ));
-    }
-    metadata.extend([
-        (
-            SOURCE_VERSION_META_KEY.to_string(),
-            Some(source_version.to_string()),
-        ),
-        (
-            SOURCE_VERSION_TS_META_KEY.to_string(),
-            Some(source_ts.to_string()),
-        ),
-        (
-            REFRESHED_AT_MS_META_KEY.to_string(),
-            Some(now_ms().to_string()),
-        ),
-        (
-            VIEW_VERSION_META_KEY.to_string(),
-            Some(predicted.to_string()),
-        ),
-    ]);
-    dataset.update_schema_metadata(metadata).await?;
+    dataset
+        .update_schema_metadata([
+            (INCARNATION_META_KEY.to_string(), Some(incarnation)),
+            (
+                SOURCE_VERSION_META_KEY.to_string(),
+                Some(source_version.to_string()),
+            ),
+            (
+                SOURCE_VERSION_TS_META_KEY.to_string(),
+                Some(source_ts.to_string()),
+            ),
+            (
+                REFRESHED_AT_MS_META_KEY.to_string(),
+                Some(now_ms().to_string()),
+            ),
+            (
+                VIEW_VERSION_META_KEY.to_string(),
+                Some(predicted.to_string()),
+            ),
+        ])
+        .await?;
    let actual = dataset.version().version;
    if actual != predicted {
        return Err(Error::Runtime {
@@ -1621,106 +1585,6 @@ mod tests {
        assert_eq!(read(view.table(), "x").await, vec![20, 40]);
    }

-    #[tokio::test]
-    async fn test_mixed_case_filter_is_canonicalized_for_lineage_and_refresh() {
-        let conn = connect("memory://").execute().await.unwrap();
-        let batch = record_batch!(
-            ("id", Int32, [1, 2, 3]),
-            ("PartyAbbrev", Utf8, ["D", "R", "D"])
-        )
-        .unwrap();
-        conn.create_table("src", batch)
-            .write_options(crate::materialized_view::tests::stable_row_ids())
-            .execute()
-            .await
-            .unwrap();
-        conn.create_materialized_view("democrats", "src")
-            .select([("id", "id")])
-            .only_if(r#""PartyAbbrev" = 'D'"#)
-            .execute()
-            .await
-            .unwrap();
-
-        // Reopen from schema metadata so these assertions cover the stored
-        // predicate and lineage, not only the declaration-time handle.
-        let view = conn.open_materialized_view("democrats").await.unwrap();
-        assert_eq!(
-            view.definition().filter.as_deref(),
-            Some("`PartyAbbrev` = 'D'")
-        );
-        assert_eq!(view.definition().inputs, ["PartyAbbrev", "id"]);
-
-        let result = view.refresh().execute().await.unwrap();
-        assert_eq!(result.rows_written, 2);
-        assert_eq!(read(view.table(), "id").await, vec![1, 3]);
-    }
-
-    #[tokio::test]
-    async fn test_legacy_raw_filter_rebuilds_and_persists_canonical_definition() {
-        let conn = connect("memory://").execute().await.unwrap();
-        let batch = record_batch!(
-            ("id", Int32, [1, 2, 3]),
-            ("PartyAbbrev", Utf8, ["D", "R", "D"])
-        )
-        .unwrap();
-        conn.create_table("legacy_src", batch)
-            .write_options(crate::materialized_view::tests::stable_row_ids())
-            .execute()
-            .await
-            .unwrap();
-        let view = conn
-            .create_materialized_view("legacy_view", "legacy_src")
-            .select([("id", "id")])
-            .only_if(r#""PartyAbbrev" = 'X'"#)
-            .execute()
-            .await
-            .unwrap();
-        assert_eq!(view.refresh().execute().await.unwrap().rows_written, 0);
-
-        // Model a definition and up-to-date watermark written before filter
-        // canonicalization was applied to materialized views.
-        let mut legacy = view.definition().clone();
-        legacy.filter = Some(r#""PartyAbbrev" = 'D'"#.into());
-        legacy.inputs = vec!["id".into()];
-        let native = view.table().as_native().unwrap();
-        let mut dataset = native.dataset.get().await.unwrap().as_ref().clone();
-        let predicted = dataset.version().version + 1;
-        dataset
-            .update_schema_metadata([
-                (
-                    DEFINITION_META_KEY.to_string(),
-                    Some(definition_to_metadata(&legacy).unwrap()),
-                ),
-                (
-                    VIEW_VERSION_META_KEY.to_string(),
-                    Some(predicted.to_string()),
-                ),
-            ])
-            .await
-            .unwrap();
-        native.dataset.update(dataset);
-
-        let reopened = conn.open_materialized_view("legacy_view").await.unwrap();
-        let result = reopened.refresh().execute().await.unwrap();
-        assert_eq!(result.mode, RefreshMode::Rebuild);
-        assert_eq!(result.rows_written, 2);
-        assert_eq!(read(reopened.table(), "id").await, vec![1, 3]);
-
-        // A fresh handle proves the migration was stored alongside the new
-        // watermark and therefore happens only once.
-        let migrated = conn.open_materialized_view("legacy_view").await.unwrap();
-        assert_eq!(
-            migrated.definition().filter.as_deref(),
-            Some("`PartyAbbrev` = 'D'")
-        );
-        assert_eq!(migrated.definition().inputs, ["PartyAbbrev", "id"]);
-        assert_eq!(
-            migrated.refresh().execute().await.unwrap().mode,
-            RefreshMode::NoOp
-        );
-        assert_eq!(read(migrated.table(), "id").await, vec![1, 3]);
-    }
-
    #[tokio::test]
    async fn test_append_refreshes_incrementally() {
        let (_conn, source, view) = refreshed_doubled(vec![1, 2]).await;
@@ -2903,7 +2767,7 @@ mod tests {
        let stale = view_native.dataset.get().await.unwrap().as_ref().clone();
        view.table().delete("x = 1").await.unwrap();

-        let err = stamp_watermark(view_native, stale, 99, 99, None, None).await;
+        let err = stamp_watermark(view_native, stale, 99, 99, None).await;
        assert!(err.is_err());

        let result = view.refresh().execute().await.unwrap();
@@ -399,9 +399,6 @@ pub trait QueryBase {
    /// x > 5 OR y = 'test'
    /// ```
    ///
-    /// Identifiers may be delimited with SQL-standard double quotes or
-    /// backticks. String literals must use single quotes.
-    ///
    /// Filtering performance can often be improved by creating a scalar index
    /// on the filter column(s).
    ///
@@ -916,17 +913,6 @@ impl QueryRequest {
    /// use different representations) the error is recorded and surfaced later
    /// by [`Self::check_filter`].
    pub(crate) fn add_filter(&mut self, new: QueryFilter) {
-        let new = match new {
-            QueryFilter::Sql(filter) => match crate::expr::canonicalize_sql_predicate(&filter) {
-                Ok(filter) => QueryFilter::Sql(filter),
-                Err(err) => {
-                    self.filter_error = Some(err.to_string());
-                    return;
-                }
-            },
-            other => other,
-        };
-
        self.filter = Some(match self.filter.take() {
            None => new,
            Some(existing) => match and_filters(existing, new) {
@@ -1188,12 +1174,12 @@ impl VectorQuery {

    /// Add another query vector to the search.
    ///
-    /// Multiple searches will be dispatched as a batch. Flat searches share
-    /// one table scan across the query vectors, avoiding the scan and memory
-    /// amplification of issuing the searches concurrently. Indexed searches
-    /// may still perform per-vector index work.
+    /// Multiple searches will be dispatched as part of the query.
+    /// This is a convenience method for adding multiple query vectors
+    /// to the search. It is not expected to be faster than issuing
+    /// multiple queries concurrently.
    ///
-    /// The output data will contain an additional column `query_index` which
+    /// The output data will contain an additional columns `query_index` which
    /// will contain the index of the query vector that was used to generate the
    /// result.
    pub fn add_query_vector(mut self, vector: impl IntoQueryVector) -> Result<Self> {
@@ -1660,14 +1646,10 @@ mod tests {
    use std::{collections::HashSet, sync::Arc};

    use super::*;
-    use arrow::{
-        array::downcast_array,
-        compute::concat_batches,
-        datatypes::{Int32Type, UInt8Type},
-    };
+    use arrow::{array::downcast_array, compute::concat_batches, datatypes::Int32Type};
    use arrow_array::{
-        FixedSizeListArray, Float32Array, Int32Array, RecordBatch, RecordBatchIterator,
-        StringArray, cast::AsArray, types::Float32Type,
+        FixedSizeListArray, Float32Array, Int32Array, RecordBatch, StringArray, cast::AsArray,
+        types::Float32Type,
    };
    use arrow_schema::{DataType, Field as ArrowField, Schema as ArrowSchema};
    use futures::{StreamExt, TryStreamExt};
@@ -1896,157 +1878,6 @@ mod tests {
        query.execute().await.unwrap();
    }

-    #[tokio::test]
-    async fn test_double_quoted_predicates_across_table_operations() {
-        let tmp_dir = tempdir().unwrap();
-        let dataset_path = tmp_dir.path().join("test.lance");
-        let uri = dataset_path.to_str().unwrap();
-        let schema = Arc::new(ArrowSchema::new(vec![
-            ArrowField::new("id", DataType::Int32, false),
-            ArrowField::new("PartyAbbrev", DataType::Utf8, false),
-            ArrowField::new("path", DataType::Utf8, false),
-        ]));
-        let batch = RecordBatch::try_new(
-            schema.clone(),
-            vec![
-                Arc::new(Int32Array::from(vec![1, 2, 3, 4])),
-                Arc::new(StringArray::from(vec!["D", "R", "R", "D"])),
-                Arc::new(StringArray::from(vec!["\\", "\\", "x", "x"])),
-            ],
-        )
-        .unwrap();
-
-        let conn = connect(uri).execute().await.unwrap();
-        let table = conn.create_table("parties", batch).execute().await.unwrap();
-        let batches = table
-            .query()
-            .only_if(r#""PartyAbbrev" = 'D'"#)
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-
-        assert_eq!(batches.iter().map(RecordBatch::num_rows).sum::<usize>(), 2);
-        assert_eq!(
-            table
-                .count_rows(Some(r#""PartyAbbrev" = 'D'"#.to_string()))
-                .await
-                .unwrap(),
-            2
-        );
-
-        // Public BaseTable dispatch cannot bypass canonicalization.
-        let query = AnyQuery::Query(QueryRequest {
-            filter: Some(QueryFilter::Sql(r#""PartyAbbrev" = 'D'"#.to_string())),
-            ..Default::default()
-        });
-        let batches = table
-            .base_table()
-            .query(&query, Default::default())
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        assert_eq!(batches.iter().map(RecordBatch::num_rows).sum::<usize>(), 2);
-        assert_eq!(
-            table
-                .base_table()
-                .count_rows(Some(crate::table::Filter::Sql(
-                    r#""PartyAbbrev" = 'D'"#.to_string(),
-                )))
-                .await
-                .unwrap(),
-            2
-        );
-
-        for predicate in [
-            r#"id = 1 -- unmatched " in a valid SQL comment"#,
-            r#"id = 1 /* unmatched " in a valid SQL comment */"#,
-            r#"id = 1 /*! OR "PartyAbbrev" = 'D' */"#,
-            r#"path = '\' AND "PartyAbbrev" = 'D'"#,
-        ] {
-            let batches = table
-                .query()
-                .only_if(predicate)
-                .execute()
-                .await
-                .unwrap()
-                .try_collect::<Vec<_>>()
-                .await
-                .unwrap();
-            assert_eq!(batches.iter().map(RecordBatch::num_rows).sum::<usize>(), 1);
-        }
-
-        // The same canonical predicate contract applies to both merge filters.
-        let source = RecordBatch::try_new(
-            schema.clone(),
-            vec![
-                Arc::new(Int32Array::from(vec![1, 2, 3])),
-                Arc::new(StringArray::from(vec!["D", "R", "R"])),
-                Arc::new(StringArray::from(vec!["\\", "\\", "x"])),
-            ],
-        )
-        .unwrap();
-        let mut merge = table.merge_insert(&["id"]);
-        merge.when_not_matched_by_source_delete(Some(r#""PartyAbbrev" = 'D'"#.to_string()));
-        let result = table
-            .base_table()
-            .merge_insert(
-                merge,
-                Box::new(RecordBatchIterator::new(vec![Ok(source)], schema.clone())),
-            )
-            .await
-            .unwrap();
-        assert_eq!(result.num_deleted_rows, 1);
-
-        let source = RecordBatch::try_new(
-            schema.clone(),
-            vec![
-                Arc::new(Int32Array::from(vec![1, 2, 3])),
-                Arc::new(StringArray::from(vec!["U", "U", "U"])),
-                Arc::new(StringArray::from(vec!["\\", "\\", "x"])),
-            ],
-        )
-        .unwrap();
-        let mut merge = table.merge_insert(&["id"]);
-        merge.when_matched_update_all(Some(r#"target."PartyAbbrev" = 'D'"#.to_string()));
-        merge
-            .execute(Box::new(RecordBatchIterator::new(vec![Ok(source)], schema)))
-            .await
-            .unwrap();
-        assert_eq!(
-            table
-                .count_rows(Some(r#""PartyAbbrev" = 'U'"#.to_string()))
-                .await
-                .unwrap(),
-            1
-        );
-
-        let update = table
-            .update()
-            .only_if(r#""PartyAbbrev" = 'R'"#)
-            .column("PartyAbbrev", "'X'");
-        table.base_table().update(update).await.unwrap();
-        assert_eq!(
-            table
-                .count_rows(Some(r#""PartyAbbrev" = 'X'"#.to_string()))
-                .await
-                .unwrap(),
-            2
-        );
-
-        let result = table
-            .base_table()
-            .delete(crate::table::Predicate::String(r#""PartyAbbrev" = 'X'"#))
-            .await
-            .unwrap();
-        assert_eq!(result.num_deleted_rows, 2);
-        assert_eq!(table.count_rows(None).await.unwrap(), 1);
-    }
-
    #[tokio::test]
    async fn test_select_with_transform() {
        let batches = make_non_empty_batches();
@@ -2503,8 +2334,7 @@ mod tests {
            .limit(1);

        let plan = query.explain_plan(true).await.unwrap();
-        assert!(plan.contains("KNNVectorDistance: queries=2"));
-        assert!(!plan.contains("UnionExec"));
+        assert!(plan.contains("UnionExec"));

        let results = query
            .execute()
@@ -2519,100 +2349,6 @@ mod tests {
        // We don't guarantee order.
        assert!(query_index.values().contains(&0));
        assert!(query_index.values().contains(&1));
-
-        // Batch KNN does not support a per-query offset, so offset queries keep
-        // the legacy per-vector plan to preserve their result semantics.
-        let offset_query = table
-            .query()
-            .nearest_to(&[0.1, 0.2, 0.3, 0.4])
-            .unwrap()
-            .add_query_vector(&[0.5, 0.6, 0.7, 0.8])
-            .unwrap()
-            .limit(1)
-            .offset(1);
-        assert!(
-            offset_query
-                .explain_plan(true)
-                .await
-                .unwrap()
-                .contains("UnionExec")
-        );
-        let offset_results = offset_query
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        assert_eq!(
-            offset_results
-                .iter()
-                .map(RecordBatch::num_rows)
-                .sum::<usize>(),
-            2
-        );
-    }
-
-    #[tokio::test]
-    async fn test_multiple_binary_query_vectors() {
-        let vectors = FixedSizeListArray::from_iter_primitive::<UInt8Type, _, _>(
-            vec![
-                Some(vec![Some(0), Some(0)]),
-                Some(vec![Some(255), Some(255)]),
-            ],
-            2,
-        );
-        let schema = Arc::new(ArrowSchema::new(vec![
-            ArrowField::new("id", DataType::Int32, false),
-            ArrowField::new("vector", vectors.data_type().clone(), false),
-        ]));
-        let batch = RecordBatch::try_new(
-            schema,
-            vec![Arc::new(Int32Array::from(vec![0, 1])), Arc::new(vectors)],
-        )
-        .unwrap();
-
-        let conn = connect("memory://").execute().await.unwrap();
-        let table = conn
-            .create_table("binary_batch", batch)
-            .execute()
-            .await
-            .unwrap();
-        let query = table
-            .query()
-            .nearest_to(&[0.0, 0.0])
-            .unwrap()
-            .add_query_vector(&[255.0, 255.0])
-            .unwrap()
-            .distance_type(DistanceType::Hamming)
-            .limit(1);
-
-        // Binary queries retain the per-vector plan because Lance's binary
-        // nearest path requires primitive UInt8 query arrays.
-        assert!(
-            query
-                .explain_plan(true)
-                .await
-                .unwrap()
-                .contains("UnionExec")
-        );
-
-        let results = query
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let results = concat_batches(&results[0].schema(), &results).unwrap();
-        assert_eq!(results.num_rows(), 2);
-
-        let ids = results["id"].as_primitive::<Int32Type>();
-        assert!(ids.values().contains(&0));
-        assert!(ids.values().contains(&1));
-        let query_index = results["query_index"].as_primitive::<Int32Type>();
-        assert!(query_index.values().contains(&0));
-        assert!(query_index.values().contains(&1));
    }

    #[tokio::test]
@@ -87,10 +87,6 @@ impl ServerVersion {
    pub fn support_blobs(&self) -> bool {
        self.0 >= semver::Version::new(0, 5, 0)
    }
-
-    pub fn support_fts_document_granularity(&self) -> bool {
-        self.0 >= semver::Version::new(0, 6, 0)
-    }
 }

 pub const OPT_REMOTE_PREFIX: &str = "remote_database_";
@@ -6,7 +6,6 @@
 use std::ops::Range;
 use std::sync::Arc;
 use std::sync::atomic::{AtomicBool, Ordering};
-use std::time::Duration;

 use arrow_array::{Array, LargeBinaryArray};
 use arrow_schema::DataType;
@@ -21,7 +20,7 @@ use crate::error::Result;
 use crate::remote::client::{HttpSend, RequestResultExt, RestfulLanceDbClient};
 use crate::table::BaseTable;

-use super::{FreshnessHeaders, FreshnessState, RemoteTable, freshness_headers_snapshot};
+use super::{FreshnessHeaders, RemoteTable};

 #[derive(Debug, Clone, Copy)]
 enum RangeRequestMode {
@@ -44,10 +43,7 @@ struct TableBlobRangeRequester<S: HttpSend> {
    path: String,
    version: Option<u64>,
    branch: Option<String>,
-    freshness: Arc<std::sync::Mutex<FreshnessState>>,
-    parent_freshness: Arc<std::sync::Mutex<FreshnessState>>,
-    parent_freshness_request: FreshnessHeaders,
-    read_consistency_interval: Option<Duration>,
+    freshness: FreshnessHeaders,
 }

 #[async_trait::async_trait]
@@ -57,9 +53,8 @@ impl<S: HttpSend> BlobRangeRequester for TableBlobRangeRequester<S> {
        range_header: &str,
        mode: RangeRequestMode,
    ) -> Result<(String, Response)> {
-        let freshness_request =
-            freshness_headers_snapshot(&self.freshness, self.read_consistency_interval);
-        let mut request = freshness_request
+        let mut request = self
+            .freshness
            .apply(self.client.get(&self.path))
            .header(header::RANGE, range_header);
        if let Some(version) = self.version {
@@ -76,9 +71,6 @@ impl<S: HttpSend> BlobRangeRequester for TableBlobRangeRequester<S> {
            return Ok((request_id, response));
        }
        let response = self.client.check_response(&request_id, response).await?;
-        freshness_request.observe_headers(&self.freshness, response.headers());
-        self.parent_freshness_request
-            .observe_headers(&self.parent_freshness, response.headers());
        Ok((request_id, response))
    }
 }
@@ -369,21 +361,18 @@ impl<S: HttpSend> RemoteTable<S> {
                message: "fetch_blobs is not supported on this LanceDB Cloud server".into(),
            });
        }
-        let read_snapshot = self.snapshot_read_state().await;
+        let version = self.current_version().await;
        let mut body = serde_json::json!({
-            "version": read_snapshot.version,
+            "version": version,
            "column": column,
            "row_ids": row_ids,
        });
        self.apply_branch_body(&mut body);

        let request = self
-            .client
-            .post(&format!("/v1/table/{}/fetch_blobs/", self.identifier))
+            .post_read(&format!("/v1/table/{}/fetch_blobs/", self.identifier))
            .json(&body);
-        let (request_id, response) = self
-            .send_with_freshness(request, true, read_snapshot.freshness)
-            .await?;
+        let (request_id, response) = self.send(request, true).await?;
        let mut stream = self.read_arrow_response(&request_id, response).await?;

        let mut blob_chunks: Vec<Arc<dyn Array>> = Vec::new();
@@ -459,7 +448,8 @@ impl<S: HttpSend> RemoteTable<S> {
            });
        }

-        let read_snapshot = self.snapshot_read_state().await;
+        let version = self.current_version().await;
+        let freshness = self.snapshot_freshness_headers();
        let encoded_column = urlencoding::encode(column);
        let requesters = row_ids
            .iter()
@@ -471,12 +461,9 @@ impl<S: HttpSend> RemoteTable<S> {
                let requester: Arc<dyn BlobRangeRequester> = Arc::new(TableBlobRangeRequester {
                    client: self.client.clone(),
                    path,
-                    version: read_snapshot.version,
+                    version,
                    branch: self.branch.clone(),
-                    freshness: Arc::new(std::sync::Mutex::new(read_snapshot.freshness_state)),
-                    parent_freshness: self.freshness.clone(),
-                    parent_freshness_request: read_snapshot.freshness,
-                    read_consistency_interval: self.client.read_consistency_interval,
+                    freshness,
                });
                requester
            })
@@ -698,46 +685,6 @@ mod tests {
        assert!(requests.lock().unwrap().contains(&"bytes=5-11".to_string()));
    }

-    #[tokio::test]
-    async fn remote_blob_file_keeps_the_open_timeline_after_parent_checkout() {
-        let range_requests = Arc::new(StdMutex::new(Vec::new()));
-        let captured = range_requests.clone();
-        let table = RemoteTable::new_mock(
-            "my_table".to_string(),
-            move |request| match request.url().path() {
-                "/v1/table/my_table/describe/" => http::Response::builder()
-                    .status(200)
-                    .body(r#"{"version":5,"schema":{"fields":[]}}"#.as_bytes().to_vec())
-                    .unwrap(),
-                "/v1/table/my_table/blob/image/10/bytes" => {
-                    captured.lock().unwrap().push((
-                        request.url().query().unwrap_or_default().to_string(),
-                        request.headers().clone(),
-                    ));
-                    range_response(&request, PAYLOAD)
-                }
-                path => panic!("unexpected path: {path}"),
-            },
-            Some(Version::new(0, 5, 0)),
-        );
-
-        table.checkout(5).await.unwrap();
-        let file = table
-            .fetch_blob_files_impl("image", &[10])
-            .await
-            .unwrap()
-            .pop()
-            .flatten()
-            .unwrap();
-        table.checkout_latest().await.unwrap();
-        file.read_range(5..12).await.unwrap();
-
-        let requests = range_requests.lock().unwrap();
-        let (query, headers) = requests.last().unwrap();
-        assert!(query.contains("version=5"));
-        assert!(!headers.contains_key("x-lancedb-min-timestamp"));
-    }
-
    #[tokio::test]
    async fn remote_blob_file_reuses_sequential_response_until_seek() {
        let requests = Arc::new(StdMutex::new(Vec::new()));
@@ -24,10 +24,7 @@ use lance::io::exec::utils::InstrumentedRecordBatchStreamAdapter;
 use crate::Error;
 use crate::remote::ARROW_STREAM_CONTENT_TYPE;
 use crate::remote::client::{HttpSend, RestfulLanceDbClient, Sender};
-use crate::remote::table::{
-    FreshnessHeaders, FreshnessState, MergeInsertRequest, REQUEST_TIMEOUT_HEADER, RemoteTable,
-    freshness_headers_snapshot,
-};
+use crate::remote::table::{MergeInsertRequest, REQUEST_TIMEOUT_HEADER, RemoteTable};
 use crate::table::datafusion::insert::COUNT_SCHEMA;
 use crate::table::write_progress::WriteProgressTracker;
 use crate::table::{AddResult, MergeResult};
@@ -57,38 +54,6 @@ pub enum WriteResult {
    Merge(MergeResult),
 }

-#[derive(Debug, Clone, Default)]
-struct WriteFreshness {
-    state: Option<Arc<Mutex<FreshnessState>>>,
-    read_consistency_interval: Option<Duration>,
-}
-
-impl WriteFreshness {
-    fn prepare(
-        &self,
-        request: reqwest::RequestBuilder,
-    ) -> (reqwest::RequestBuilder, Option<FreshnessHeaders>) {
-        match &self.state {
-            Some(state) => {
-                let freshness_request =
-                    freshness_headers_snapshot(state, self.read_consistency_interval);
-                (freshness_request.apply(request), Some(freshness_request))
-            }
-            None => (request, None),
-        }
-    }
-
-    fn observe(
-        &self,
-        freshness_request: Option<FreshnessHeaders>,
-        headers: &reqwest::header::HeaderMap,
-    ) {
-        if let (Some(state), Some(freshness_request)) = (&self.state, freshness_request) {
-            freshness_request.observe_headers(state, headers);
-        }
-    }
-}
-
 /// ExecutionPlan for streaming a write (add or merge_insert) to a remote
 /// LanceDB table.
 ///
@@ -106,7 +71,6 @@ pub struct RemoteWriteExec<S: HttpSend = Sender> {
    table_name: String,
    identifier: String,
    client: RestfulLanceDbClient<S>,
-    freshness: WriteFreshness,
    input: Arc<dyn ExecutionPlan>,
    op: WriteOp,
    properties: Arc<PlanProperties>,
@@ -206,7 +170,6 @@ impl<S: HttpSend + 'static> RemoteWriteExec<S> {
            table_name,
            identifier,
            client,
-            freshness: WriteFreshness::default(),
            input,
            op,
            properties: Arc::new(properties),
@@ -220,18 +183,6 @@ impl<S: HttpSend + 'static> RemoteWriteExec<S> {
        }
    }

-    pub(super) fn with_freshness(
-        mut self,
-        state: Arc<Mutex<FreshnessState>>,
-        read_consistency_interval: Option<Duration>,
-    ) -> Self {
-        self.freshness = WriteFreshness {
-            state: Some(state),
-            read_consistency_interval,
-        };
-        self
-    }
-
    /// Get the add result after execution, if this exec ran an insert.
    pub fn add_result(&self) -> Option<AddResult> {
        match self
@@ -334,7 +285,6 @@ impl<S: HttpSend + 'static> RemoteWriteExec<S> {
 /// each threading the same handful of arguments.
 struct PartRequestCtx<'a, S: HttpSend> {
    client: &'a RestfulLanceDbClient<S>,
-    freshness: &'a WriteFreshness,
    identifier: &'a str,
    table_name: &'a str,
    upload_id: &'a str,
@@ -402,11 +352,7 @@ impl<S: HttpSend + 'static> PartRequestCtx<'_, S> {
    }

    /// Build the `/insert` request for a single multipart part.
-    fn build_part_request(
-        &self,
-        part_id: &str,
-        body: reqwest::Body,
-    ) -> (reqwest::RequestBuilder, Option<FreshnessHeaders>) {
+    fn build_part_request(&self, part_id: &str, body: reqwest::Body) -> reqwest::RequestBuilder {
        let mut request = self
            .client
            .post(&format!("/v1/table/{}/insert/", self.identifier))
@@ -422,16 +368,12 @@ impl<S: HttpSend + 'static> PartRequestCtx<'_, S> {
        if let Some(b) = self.branch {
            request = request.query(&[("branch", b)]);
        }
-        self.freshness.prepare(request.body(body))
+        request.body(body)
    }

    /// Send a single part's request and drain the response, mapping HTTP and
    /// table-not-found errors into `DataFusionError`.
-    async fn send_part_request(
-        &self,
-        request: reqwest::RequestBuilder,
-        freshness_request: Option<FreshnessHeaders>,
-    ) -> DataFusionResult<()> {
+    async fn send_part_request(&self, request: reqwest::RequestBuilder) -> DataFusionResult<()> {
        let (request_id, response) = self
            .client
            .send(request)
@@ -446,8 +388,6 @@ impl<S: HttpSend + 'static> PartRequestCtx<'_, S> {
            .check_response(&request_id, response)
            .await
            .map_err(|e| DataFusionError::External(Box::new(e)))?;
-        self.freshness
-            .observe(freshness_request, response.headers());
        response.bytes().await.map_err(|e| {
            DataFusionError::External(Box::new(Error::Http {
                source: Box::new(e),
@@ -479,7 +419,7 @@ impl<S: HttpSend + 'static> PartRequestCtx<'_, S> {
        let body = reqwest::Body::wrap_stream(chunk_rx);

        let part_id = uuid::Uuid::new_v4().to_string();
-        let (request, freshness_request) = self.build_part_request(&part_id, body);
+        let request = self.build_part_request(&part_id, body);

        // Measured from just before the request is sent, matching the window the
        // client read timeout applies to the upload.
@@ -555,7 +495,7 @@ impl<S: HttpSend + 'static> PartRequestCtx<'_, S> {
            Ok::<bool, DataFusionError>(input_ended)
        };

-        let send = self.send_part_request(request, freshness_request);
+        let send = self.send_part_request(request);

        // `join!` rather than `tokio::spawn`: the producer borrows `input` (and
        // `schema`), so it cannot satisfy the `'static` bound a spawned task
@@ -629,7 +569,7 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {
        // Building a fresh exec (with a new, empty `result`) is what makes the
        // outer rescannable retry loop work: `reset_state()` clears the captured
        // result so a re-execution starts clean.
-        let mut exec = Self::new_inner(
+        Ok(Arc::new(Self::new_inner(
            self.table_name.clone(),
            self.identifier.clone(),
            self.client.clone(),
@@ -640,9 +580,7 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {
            self.branch.clone(),
            self.max_bytes_per_request,
            self.max_request_duration,
-        );
-        exec.freshness = self.freshness.clone();
-        Ok(Arc::new(exec))
+        )))
    }

    fn execute(
@@ -675,7 +613,6 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {
                &self.metrics,
            ));
        let client = self.client.clone();
-        let freshness = self.freshness.clone();
        let identifier = self.identifier.clone();
        let op = self.op.clone();
        let result_slot = self.result.clone();
@@ -697,7 +634,6 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {
                let overwrite = matches!(op, WriteOp::Insert { overwrite: true });
                let ctx = PartRequestCtx {
                    client: &client,
-                    freshness: &freshness,
                    identifier: &identifier,
                    table_name: &table_name,
                    upload_id,
@@ -752,7 +688,7 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {

            let (error_tx, mut error_rx) = tokio::sync::oneshot::channel();
            let body = Self::stream_as_http_body(input_stream, error_tx, tracker)?;
-            let (request, freshness_request) = freshness.prepare(request.body(body));
+            let request = request.body(body);

            let result: DataFusionResult<(String, _)> = async {
                let (request_id, response) = client
@@ -772,7 +708,6 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteWriteExec<S> {
                    .check_response(&request_id, response)
                    .await
                    .map_err(|e| DataFusionError::External(Box::new(e)))?;
-                freshness.observe(freshness_request, response.headers());

                Ok((request_id, response))
            }
@@ -59,9 +59,7 @@ use crate::index::{IndexConfig, IndexStatisticsImpl, IndexType};
 use crate::job::Job;
 use crate::query::{IntoQueryVector, Query, QueryExecutionOptions, TakeQuery, VectorQuery};
 use crate::table::datafusion::insert::InsertExec;
-use crate::utils::{
-    PatchReadParam, PatchWriteParam, public_fts_field_path_by_id, resolve_arrow_field_path,
-};
+use crate::utils::{PatchReadParam, PatchWriteParam, resolve_arrow_field_path};

 use self::dataset::DatasetConsistencyWrapper;
 use self::merge::MergeInsertBuilder;
@@ -103,9 +101,7 @@ pub use lance::dataset::refs::{BranchContents, Ref, TagContents, Tags as LanceTa
 pub use lance::dataset::scanner::DatasetRecordBatchStream;
 pub use lance_index::optimize::OptimizeOptions;
 pub use lsm_stats::{BucketStats, GenerationStats, LsmStats, MemtableStats};
-pub use optimize::{
-    CompactionMode, CompactionOptions, IndexRemapMode, OptimizeAction, OptimizeStats,
-};
+pub use optimize::{CompactionOptions, OptimizeAction, OptimizeStats};
 pub use refresh::RefreshColumnResult;
 pub use schema_evolution::{
    AddColumnsResult, AlterColumnsResult, DropColumnsResult, FieldMetadataUpdate,
@@ -564,13 +560,6 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
    fn id(&self) -> &str;
    /// Get the arrow [Schema] of the table.
    async fn schema(&self) -> Result<SchemaRef>;
-    /// Create a read-only handle pinned to the table's current active revision.
-    ///
-    /// The returned handle is independent from later refreshes or checkouts on
-    /// this handle.  This is used by bindings that must prepare client-side
-    /// query state from the same revision that the query will execute against.
-    #[doc(hidden)]
-    async fn query_snapshot(&self) -> Result<Arc<dyn BaseTable>>;
    /// Count the number of rows in this table.
    async fn count_rows(&self, filter: Option<Filter>) -> Result<usize>;
    /// Create a physical plan for the query.
@@ -752,8 +741,8 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
    /// Declare computed columns, each defined by a SQL expression.
    ///
    /// Where the declaration is planned depends on the backend: a local table
-    /// validates and types the expression itself, while a remote one sends the
-    /// expression for the server to plan.
+    /// validates and types the expression itself, a remote one sends the text
+    /// for the server to plan.
    async fn add_computed_columns(
        &self,
        _columns: &[(String, String)],
@@ -796,12 +785,6 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
    async fn drop_columns(&self, columns: &[&str]) -> Result<DropColumnsResult>;
    /// Get the version of the table.
    async fn version(&self) -> Result<u64>;
-    /// Return a new table handle pinned to the exact revision currently visible.
-    async fn checkout_current(&self) -> Result<Arc<dyn BaseTable>> {
-        Err(Error::NotSupported {
-            message: "checkout_current is not supported on this table type".into(),
-        })
-    }
    /// Checkout a specific version of the table.
    async fn checkout(&self, version: u64) -> Result<()>;
    /// Checkout a table version referenced by a tag.
@@ -1150,26 +1133,13 @@ impl Table {
        self.inner.schema().await
    }

-    /// Create a read-only handle pinned to the current active revision.
-    #[doc(hidden)]
-    pub async fn query_snapshot(&self) -> Result<Self> {
-        Ok(Self {
-            inner: self.inner.query_snapshot().await?,
-            database: self.database.clone(),
-            embedding_registry: self.embedding_registry.clone(),
-        })
-    }
-
    /// Count the number of rows in this dataset.
    ///
    /// # Arguments
    ///
    /// * `filter` if present, only count rows matching the filter
    pub async fn count_rows(&self, filter: Option<String>) -> Result<usize> {
-        let filter = filter
-            .map(|predicate| crate::expr::canonicalize_sql_predicate(&predicate).map(Filter::Sql))
-            .transpose()?;
-        self.inner.count_rows(filter).await
+        self.inner.count_rows(filter.map(Filter::Sql)).await
    }

    /// Names of the blob v2 columns in this table, in declaration order.
@@ -1369,13 +1339,7 @@ impl Table {
    /// # });
    /// ```
    pub async fn delete(&self, predicate: impl Into<Predicate<'_>>) -> Result<DeleteResult> {
-        match predicate.into() {
-            Predicate::String(predicate) => {
-                let predicate = crate::expr::canonicalize_sql_predicate(predicate)?;
-                self.inner.delete(Predicate::String(&predicate)).await
-            }
-            predicate @ Predicate::Expr(_) => self.inner.delete(predicate).await,
-        }
+        self.inner.delete(predicate.into()).await
    }

    /// Create an index on the provided column(s).
@@ -1788,23 +1752,7 @@ impl Table {
        self.inner.alter_columns(alterations).await
    }

-    /// Update per-field (column) metadata.
-    ///
-    /// Each [`FieldMetadataUpdate`] is merged into the field's existing metadata
-    /// by default; use [`FieldMetadataUpdate::remove`] to delete a key, or
-    /// [`FieldMetadataUpdate::replace`] to swap the field's entire metadata map.
-    ///
-    /// The following keys are treated specially, by convention, and should be
-    /// used when appropriate:
-    ///
-    /// - `lancedb:description`: for a human-readable description of a field.
-    /// - `lancedb:tag:<name>`: for a user-defined key-value tag, where the suffix
-    ///   names the tag category; e.g. `lancedb:tag:model: "clip"`.
-    /// - `lancedb:logical-column`: for a column grouping; e.g. `feature_v1` and
-    ///   `feature_v2` might be in the same logical column.
-    /// - `lancedb:status`: for status options (`production`, `candidate`,
-    ///   `deprecated`, `archived`) to designate the current life cycle state of
-    ///   this column.
+    /// Update per-field metadata (merges by default).
    pub async fn update_field_metadata(
        &self,
        updates: &[FieldMetadataUpdate],
@@ -1996,20 +1944,6 @@ impl Table {
        self.inner.version().await
    }

-    /// Return a new table handle pinned to the exact revision currently visible.
-    ///
-    /// This is used when asynchronous preparation must remain consistent with
-    /// the revision used for a later read.
-    #[doc(hidden)]
-    pub async fn checkout_current(&self) -> Result<Self> {
-        let inner = self.inner.checkout_current().await?;
-        Ok(Self {
-            inner,
-            database: self.database.clone(),
-            embedding_registry: self.embedding_registry.clone(),
-        })
-    }
-
    /// Checks out a specific version of the Table
    ///
    /// Any read operation on the table will now access the data at the checked out version.
@@ -3105,33 +3039,10 @@ impl BaseTable for NativeTable {
        &self.id
    }

-    async fn query_snapshot(&self) -> Result<Arc<dyn BaseTable>> {
-        let snapshot = self.dataset.new_query_snapshot().await?;
-        let mut table = self.with_dataset(snapshot);
-        // QueryTable requests do not carry a revision. A pinned snapshot must
-        // execute locally until the namespace API can accept that revision.
-        table
-            .pushdown_operations
-            .remove(&NamespaceClientPushdownOperation::QueryTable);
-        Ok(Arc::new(table))
-    }
-
    async fn version(&self) -> Result<u64> {
        Ok(self.dataset.get().await?.version().version)
    }

-    async fn checkout_current(&self) -> Result<Arc<dyn BaseTable>> {
-        let current = self.dataset.get().await?;
-        let dataset = dataset::DatasetConsistencyWrapper::new_time_travel(
-            current.as_ref().clone(),
-            self.read_consistency_interval,
-        );
-        Ok(Arc::new(Self {
-            dataset,
-            ..self.clone()
-        }))
-    }
-
    async fn checkout(&self, version: u64) -> Result<()> {
        self.dataset.as_time_travel(version).await
    }
@@ -3250,10 +3161,7 @@ impl BaseTable for NativeTable {
        let dataset = self.dataset.get().await?;
        match filter {
            None => Ok(dataset.count_rows(None).await?),
-            Some(Filter::Sql(sql)) => {
-                let sql = crate::expr::canonicalize_sql_predicate(&sql)?;
-                Ok(dataset.count_rows(Some(sql)).await?)
-            }
+            Some(Filter::Sql(sql)) => Ok(dataset.count_rows(Some(sql)).await?),
            Some(Filter::Datafusion(_)) => Err(Error::NotSupported {
                message: "Datafusion filters are not yet supported".to_string(),
            }),
@@ -3300,7 +3208,7 @@ impl BaseTable for NativeTable {

        let output = add.into_plan(&table_schema, &table_def)?;

-        let mut lance_params = output
+        let lance_params = output
            .write_options
            .lance_write_params
            .unwrap_or(WriteParams {
@@ -3310,9 +3218,6 @@ impl BaseTable for NativeTable {
                },
                ..Default::default()
            });
-        if output.allow_external_blob_outside_bases {
-            lance_params.allow_external_blob_outside_bases = true;
-        }

        // Repartition for write parallelism if beneficial.
        let plan = if num_partitions > 1 {
@@ -3591,14 +3496,7 @@ impl BaseTable for NativeTable {
                let field_ids = idx_desc.field_ids();
                let mut columns = Vec::with_capacity(field_ids.len());
                for field_id in field_ids {
-                    let field_path = match if index_type == crate::index::IndexType::FTS {
-                        public_fts_field_path_by_id(dataset.schema(), *field_id as i32)
-                    } else {
-                        dataset
-                            .schema()
-                            .field_path(*field_id as i32)
-                            .map_err(Into::into)
-                    } {
+                    let field_path = match dataset.schema().field_path(*field_id as i32) {
                        Ok(field_path) => field_path,
                        Err(e) => {
                            log::warn!(
@@ -4185,14 +4083,6 @@ mod tests {
                parent_list_calls: self.parent_list_calls.clone(),
            })
        }
-
-        fn wrap_paginated(
-            &self,
-            _store_prefix: &str,
-            _original: Arc<dyn object_store::list::PaginatedListStore>,
-        ) -> Option<Arc<dyn object_store::list::PaginatedListStore>> {
-            None
-        }
    }

    #[tokio::test]
@@ -4296,14 +4186,6 @@ mod tests {
            self.called.store(true, Ordering::Relaxed);
            original
        }
-
-        fn wrap_paginated(
-            &self,
-            _store_prefix: &str,
-            original: Arc<dyn object_store::list::PaginatedListStore>,
-        ) -> Option<Arc<dyn object_store::list::PaginatedListStore>> {
-            Some(original)
-        }
    }

    #[tokio::test]
@@ -60,7 +60,6 @@ pub struct AddDataBuilder {
    pub(crate) embedding_registry: Option<Arc<dyn EmbeddingRegistry>>,
    pub(crate) progress_callback: Option<ProgressCallback>,
    pub(crate) write_parallelism: Option<usize>,
-    pub(crate) allow_external_blob_outside_bases: bool,
 }

 impl std::fmt::Debug for AddDataBuilder {
@@ -88,7 +87,6 @@ impl AddDataBuilder {
            embedding_registry,
            progress_callback: None,
            write_parallelism: None,
-            allow_external_blob_outside_bases: false,
        }
    }

@@ -143,16 +141,6 @@ impl AddDataBuilder {
        self
    }

-    /// Store blob URIs that sit outside registered blob bases.
-    ///
-    /// The row keeps a reference, so the object has to stay readable.
-    /// [`crate::table::Table::fetch_blobs`] reads from that location.
-    /// Defaults to `false`. Local tables only.
-    pub fn allow_external_blob_outside_bases(mut self, allow: bool) -> Self {
-        self.allow_external_blob_outside_bases = allow;
-        self
-    }
-
    pub async fn execute(self) -> Result<AddResult> {
        if self.write_parallelism.map(|p| p == 0).unwrap_or(false) {
            return Err(Error::InvalidInput {
@@ -211,7 +199,6 @@ impl AddDataBuilder {
            write_options: self.write_options,
            mode: self.mode,
            tracker,
-            allow_external_blob_outside_bases: self.allow_external_blob_outside_bases,
        })
    }
 }
@@ -225,7 +212,6 @@ pub struct PreprocessingOutput {
    pub write_options: WriteOptions,
    pub mode: AddDataMode,
    pub tracker: Option<Arc<WriteProgressTracker>>,
-    pub allow_external_blob_outside_bases: bool,
 }

 /// Check that the input schema is valid for insert.
@@ -9,35 +9,29 @@
 //! refresh fills the rows.
 //!
 //! The rule is tagged by kind ([`ComputedColumnKind`]) because kinds differ in
-//! where the column's type and inputs come from. A SQL expression determines
-//! its inputs and physical result type. A direct projection of a Blob v2 field
-//! also inherits that field's semantic type while execution continues to use
-//! `LargeBinary`. A kind resolved through a registry cannot be typed without
-//! consulting it.
-//! Registered Functions use an exact remote version plus a schema-level
-//! Function binding; unknown newer kinds remain readable and fail closed
-//! before mutation.
+//! where the column's type and inputs come from. A SQL expression is
+//! self-describing -- both are derived from the expression, so a caller writes
+//! neither -- while a kind resolved through a registry cannot be typed without
+//! consulting it. Registered Functions use an exact remote version plus a
+//! schema-level Function binding; unknown newer kinds remain readable and fail
+//! closed before mutation.
 //!
 //! [`computed_columns`] and [`computed_column_from_field`] read declarations
 //! back off a schema.

-use std::collections::{BTreeSet, HashMap, HashSet};
+use std::collections::{BTreeSet, HashMap};
 use std::sync::Arc;

-use arrow_schema::{DataType, Field as ArrowField, Fields, Schema as ArrowSchema, SchemaRef};
-use datafusion_common::{ScalarValue, tree_node::TreeNode};
-use datafusion_expr::Expr;
+use arrow_schema::{DataType, Field as ArrowField, Schema as ArrowSchema, SchemaRef};
+use datafusion_common::tree_node::TreeNode;
 use datafusion_physical_plan::PhysicalExpr;
 use lance::dataset::NewColumnTransform;
-use lance_arrow::FieldExt;
-use lance_core::datatypes::{BLOB_V2_DESC_FIELD, format_field_path_minimal, parse_field_path};
 use lance_datafusion::planner::Planner;
 use lance_namespace::models::{JsonArrowDataType, JsonArrowField, JsonArrowSchema};
 use serde::{Deserialize, Serialize};
 use serde_json::Value;

 use crate::function::{FunctionApplication, FunctionBinding};
-use crate::utils::resolve_arrow_field_path;
 use crate::{Error, Result};

 /// Field metadata key marking a column as computed. The value is `"true"`.
@@ -1112,20 +1106,15 @@ pub(crate) fn ensure_no_foreign_declarations<'a>(
    fields: impl IntoIterator<Item = &'a Arc<ArrowField>>,
 ) -> Result<()> {
    for field in fields {
-        ensure_no_foreign_declaration(field)?;
-    }
-    Ok(())
-}
-
-fn ensure_no_foreign_declaration(field: &ArrowField) -> Result<()> {
-    if field.metadata().keys().any(|k| is_declaration_key(k)) {
-        return Err(Error::InvalidInput {
-            message: format!(
-                "field '{}' carries computed-column metadata; declare computed columns \
-                 with add_columns().computed()",
-                field.name()
-            ),
-        });
+        if field.metadata().keys().any(|k| is_declaration_key(k)) {
+            return Err(Error::InvalidInput {
+                message: format!(
+                    "field '{}' carries computed-column metadata; declare computed columns \
+                     with add_columns().computed()",
+                    field.name()
+                ),
+            });
+        }
    }
    Ok(())
 }
@@ -1173,154 +1162,15 @@ pub(crate) struct BoundExpression {
    /// The columns the expression names, as written; nested inputs keep
    /// their dotted path.
    pub inputs: Vec<String>,
-    /// The top-level columns evaluation reads, in physical-expression order.
-    /// A nested input appears through its root.
+    /// The top-level columns evaluation reads, in [`Self::read_schema`]
+    /// order. A nested input appears through its root.
    pub roots: Vec<String>,
+    /// The projected schema evaluation runs against.
+    pub read_schema: SchemaRef,
    /// The compiled expression.
    pub physical: Arc<dyn PhysicalExpr>,
    /// The type the expression yields.
    pub data_type: DataType,
-    /// Blob v2 leaves the scan must materialize as `LargeBinary`.
-    pub blob_paths: Vec<String>,
-    /// A directly projected Blob v2 field whose semantics the output inherits.
-    projected_blob_field: Option<ArrowField>,
-}
-
-fn is_direct_field_projection(expr: &Expr) -> bool {
-    match expr {
-        Expr::Column(_) => true,
-        Expr::ScalarFunction(function)
-            if function.name() == "get_field" && function.args.len() == 2 =>
-        {
-            is_direct_field_projection(&function.args[0])
-                && matches!(
-                    &function.args[1],
-                    Expr::Literal(ScalarValue::Utf8(Some(_)), _)
-                )
-        }
-        _ => false,
-    }
-}
-
-fn projected_blob_field(schema: &ArrowSchema, expr: &Expr) -> Result<Option<ArrowField>> {
-    if !is_direct_field_projection(expr) {
-        return Ok(None);
-    }
-    let paths = Planner::column_names_in_expr(expr);
-    let [path] = paths.as_slice() else {
-        return Ok(None);
-    };
-    let (_, field) = resolve_arrow_field_path(schema, path)?;
-    Ok(field.is_blob_v2().then_some(field))
-}
-
-fn collect_blob_paths(field: &ArrowField, parent: &[String], paths: &mut Vec<Vec<String>>) {
-    let mut path = parent.to_vec();
-    path.push(field.name().clone());
-    if field.is_blob_v2() {
-        paths.push(path);
-        return;
-    }
-    match field.data_type() {
-        DataType::Struct(children) => {
-            for child in children {
-                collect_blob_paths(child, &path, paths);
-            }
-        }
-        DataType::List(child)
-        | DataType::LargeList(child)
-        | DataType::FixedSizeList(child, _)
-        | DataType::Map(child, _) => collect_blob_paths(child, &path, paths),
-        _ => {}
-    }
-}
-
-fn schema_blob_paths(schema: &ArrowSchema) -> Vec<Vec<String>> {
-    let mut paths = Vec::new();
-    for field in schema.fields() {
-        collect_blob_paths(field, &[], &mut paths);
-    }
-    paths
-}
-
-fn transform_blob_field(
-    field: &ArrowField,
-    parent: &[String],
-    materialized: &HashSet<Vec<String>>,
-) -> ArrowField {
-    let mut path = parent.to_vec();
-    path.push(field.name().clone());
-    if field.is_blob_v2() {
-        if materialized.contains(&path) {
-            return ArrowField::new(field.name(), DataType::LargeBinary, field.is_nullable());
-        }
-        return ArrowField::new(
-            field.name(),
-            BLOB_V2_DESC_FIELD.data_type().clone(),
-            field.is_nullable(),
-        )
-        .with_metadata(BLOB_V2_DESC_FIELD.metadata().clone());
-    }
-
-    let data_type = match field.data_type() {
-        DataType::Struct(children) => DataType::Struct(
-            children
-                .iter()
-                .map(|child| Arc::new(transform_blob_field(child, &path, materialized)))
-                .collect(),
-        ),
-        DataType::List(child) => {
-            DataType::List(Arc::new(transform_blob_field(child, &path, materialized)))
-        }
-        DataType::LargeList(child) => {
-            DataType::LargeList(Arc::new(transform_blob_field(child, &path, materialized)))
-        }
-        DataType::FixedSizeList(child, size) => DataType::FixedSizeList(
-            Arc::new(transform_blob_field(child, &path, materialized)),
-            *size,
-        ),
-        DataType::Map(child, sorted) => DataType::Map(
-            Arc::new(transform_blob_field(child, &path, materialized)),
-            *sorted,
-        ),
-        _ => return field.clone(),
-    };
-    ArrowField::new(field.name(), data_type, field.is_nullable())
-        .with_metadata(field.metadata().clone())
-}
-
-fn blob_runtime_schema(schema: &ArrowSchema, materialized: &HashSet<Vec<String>>) -> SchemaRef {
-    Arc::new(ArrowSchema::new_with_metadata(
-        schema
-            .fields()
-            .iter()
-            .map(|field| Arc::new(transform_blob_field(field, &[], materialized)))
-            .collect::<Fields>(),
-        schema.metadata().clone(),
-    ))
-}
-
-fn referenced_blob_paths(schema: &ArrowSchema, inputs: &[String]) -> Result<Vec<Vec<String>>> {
-    let input_paths = inputs
-        .iter()
-        .map(|input| {
-            parse_field_path(input).map_err(|error| Error::InvalidInput {
-                message: format!("invalid computed-column input path '{input}': {error}"),
-            })
-        })
-        .collect::<Result<Vec<_>>>()?;
-    Ok(schema_blob_paths(schema)
-        .into_iter()
-        .filter(|blob_path| {
-            input_paths.iter().any(|input_path| {
-                input_path.len() <= blob_path.len()
-                    && input_path
-                        .iter()
-                        .zip(blob_path)
-                        .all(|(input, blob)| input == blob)
-            })
-        })
-        .collect())
 }

 /// Parse, resolve and compile `expression` against `schema`.
@@ -1335,18 +1185,10 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
        message,
    };

-    // Blob v2 is a semantic type whose runtime expression ABI is
-    // `LargeBinary`. Parse against that ABI first so a direct Blob reference
-    // is not mistaken for its storage descriptor struct.
-    let all_blob_paths = schema_blob_paths(schema.as_ref())
-        .into_iter()
-        .collect::<HashSet<_>>();
-    let parsing_schema = blob_runtime_schema(schema.as_ref(), &all_blob_paths);
-    let planner = Planner::new(parsing_schema);
+    let planner = Planner::new(schema.clone());
    let parsed = planner
        .parse_expr(expression)
        .map_err(|e| invalid(e.to_string()))?;
-    let projected_blob_field = projected_blob_field(schema.as_ref(), &parsed)?;

    // A declaration is evaluated more than once -- staging and writing are
    // separate passes, and a refresh years later replays the same text -- so
@@ -1376,19 +1218,13 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
    inputs.sort();
    inputs.dedup();

-    let blob_paths = referenced_blob_paths(schema.as_ref(), &inputs)?;
-    let runtime_schema = blob_runtime_schema(
-        schema.as_ref(),
-        &blob_paths.iter().cloned().collect::<HashSet<_>>(),
-    );
-
    // A nested input is recorded by its path but read through its root
    // column; Schema::index_of resolves top-level names only. Resolved here
    // rather than left to the planner so an unknown column names itself in
    // the error instead of surfacing as a plan failure.
    let mut indices = Vec::with_capacity(inputs.len());
    for input in &inputs {
-        let index = runtime_schema
+        let index = schema
            .index_of(root(input))
            .map_err(|_| invalid(format!("unknown column '{input}'")))?;
        if !indices.contains(&index) {
@@ -1401,7 +1237,7 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
    // compiles the expression has to be built on the projected schema
    // evaluation will actually read.
    let read_schema = Arc::new(
-        runtime_schema
+        schema
            .project(&indices)
            .map_err(|e| invalid(e.to_string()))?,
    );
@@ -1411,8 +1247,7 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
        .map(|field| field.name().clone())
        .collect();

-    let runtime_planner = Planner::new(runtime_schema);
-    let optimized = runtime_planner
+    let optimized = planner
        .optimize_expr(parsed)
        .map_err(|e| invalid(e.to_string()))?;
    let physical = Planner::new(read_schema.clone())
@@ -1425,16 +1260,9 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
    Ok(BoundExpression {
        inputs,
        roots,
+        read_schema,
        physical,
        data_type,
-        blob_paths: blob_paths
-            .iter()
-            .map(|path| {
-                let segments = path.iter().map(String::as_str).collect::<Vec<_>>();
-                format_field_path_minimal(&segments)
-            })
-            .collect(),
-        projected_blob_field,
    })
 }

@@ -1445,91 +1273,35 @@ pub(crate) fn bind(schema: SchemaRef, column: &str, expression: &str) -> Result<
 /// refresh time: that the expression parses, that every column it reads
 /// exists, and that the target name is free. A declaration that survives this
 /// is one a refresh can always act on.
-///
-/// Each accepted column joins the schema the next one resolves against, so a
-/// batch may declare `a` and then `b = a + 1` in one commit. Refresh order
-/// then matters, and refresh enforces it: `b` is refused while `a` still has
-/// unfilled rows.
-fn plan_declarations(schema: SchemaRef, columns: &[(String, String)]) -> Result<Vec<ArrowField>> {
+pub(crate) fn plan(schema: SchemaRef, columns: &[(String, String)]) -> Result<Vec<ArrowField>> {
    if columns.is_empty() {
        return Err(Error::InvalidInput {
            message: "at least one computed column is required".into(),
        });
    }

-    let mut schema = schema;
    let mut fields = Vec::with_capacity(columns.len());
+    let mut declared: Vec<&str> = Vec::with_capacity(columns.len());

    for (name, expression) in columns {
-        if schema.field_with_name(name).is_ok() {
-            return Err(Error::ColumnAlreadyExists {
-                name: name.to_string(),
-            });
+        if schema.field_with_name(name).is_ok() || declared.contains(&name.as_str()) {
+            return Err(Error::ColumnAlreadyExists { name: name.clone() });
        }

        let bound = bind(schema.clone(), name, expression)?;

        // Declared columns start entirely null, so nullability is a property
        // of the declaration rather than of what the expression yields.
-        let computed_metadata = computed_column_metadata(expression, &bound.inputs);
-        let field = match bound.projected_blob_field {
-            Some(source) => {
-                let mut metadata = source.metadata().clone();
-                metadata.retain(|key, _| !is_declaration_key(key));
-                metadata.extend(computed_metadata);
-                source
-                    .with_name(name)
-                    .with_nullable(true)
-                    .with_metadata(metadata)
-            }
-            None => ArrowField::new(name, bound.data_type, true).with_metadata(computed_metadata),
-        };
-        schema = Arc::new(ArrowSchema::new_with_metadata(
-            schema
-                .fields()
-                .iter()
-                .cloned()
-                .chain(std::iter::once(Arc::new(field.clone())))
-                .collect::<Fields>(),
-            schema.metadata().clone(),
-        ));
-        fields.push(field);
+        fields.push(
+            ArrowField::new(name, bound.data_type, true)
+                .with_metadata(computed_column_metadata(expression, &bound.inputs)),
+        );
+        declared.push(name);
    }

    Ok(fields)
 }

-pub(crate) fn plan(schema: SchemaRef, columns: &[(String, String)]) -> Result<Vec<ArrowField>> {
-    plan_declarations(schema, columns)
-}
-
-/// Run the schema-level checks of
-/// [`AddColumnsBuilder::computed`](super::AddColumnsBuilder::computed) against
-/// `schema` without committing: the Function-binding guard and the planning of
-/// every declaration. For callers that stage declarations behind other work
-/// and need those rejections before any of it lands.
-///
-/// Only the schema is consulted. Declaring also refuses a table with an LSM
-/// write spec or retained SSTables; that is table state, checked at commit.
-///
-/// ```
-/// # use std::sync::Arc;
-/// # use arrow_schema::{DataType, Field, Schema};
-/// use lancedb::table::computed_columns::validate_declarations;
-///
-/// let schema = Arc::new(Schema::new(vec![Field::new("x", DataType::Int32, false)]));
-/// let declarations = vec![
-///     ("a".to_string(), "x + 1".to_string()),
-///     ("b".to_string(), "a * 2".to_string()),
-/// ];
-/// assert!(validate_declarations(schema.clone(), &declarations).is_ok());
-/// assert!(validate_declarations(schema, &[("c".into(), "random()".into())]).is_err());
-/// ```
-pub fn validate_declarations(schema: SchemaRef, columns: &[(String, String)]) -> Result<()> {
-    ensure_no_function_bindings_for_mutation(schema.as_ref(), "schema evolution")?;
-    plan(schema, columns).map(drop)
-}
-
 /// Build the transform that declares `columns` against `schema`.
 ///
 /// An all-null column is how a binding with no values yet is carried into a
@@ -1541,7 +1313,7 @@ pub(crate) fn declare(
    schema: SchemaRef,
    columns: &[(String, String)],
 ) -> Result<NewColumnTransform> {
-    let fields = plan_declarations(schema, columns)?;
+    let fields = plan(schema, columns)?;
    Ok(NewColumnTransform::AllNulls(Arc::new(ArrowSchema::new(
        fields,
    ))))
@@ -1568,22 +1340,6 @@ pub(super) async fn add_foreign_kind(table: &crate::Table, name: &str, kind: &st

 #[cfg(test)]
 mod tests {
-    /// The gate's reproducer: the validator applies the same schema-level
-    /// guard declaring does, so a staging caller is refused before it commits
-    /// anything else.
-    #[test]
-    fn test_validate_declarations_matches_schema_admission_barriers() {
-        let schema = Arc::new(ArrowSchema::new_with_metadata(
-            vec![ArrowField::new("x", DataType::Int32, true)],
-            HashMap::from([(
-                FUNCTION_BINDINGS_META_KEY.to_string(),
-                "not valid binding metadata".to_string(),
-            )]),
-        ));
-        let declarations = vec![("a".to_string(), "x + 1".to_string())];
-        assert!(super::validate_declarations(schema, &declarations).is_err());
-    }
-
    #[test]
    fn output_arrow_type_grammar_matches_the_shared_golden() {
        let golden: serde_json::Value = serde_json::from_str(include_str!(
@@ -1667,44 +1423,6 @@ mod tests {
        );
    }

-    #[test]
-    fn test_direct_blob_projection_inherits_semantics() {
-        let schema = Arc::new(ArrowSchema::new(vec![crate::blob("image", false)]));
-        let fields = plan(
-            schema,
-            &[
-                ("first".to_string(), "image".to_string()),
-                ("second".to_string(), "first".to_string()),
-            ],
-        )
-        .unwrap();
-
-        for field in &fields {
-            assert!(field.is_blob_v2());
-            assert!(field.is_nullable());
-        }
-        assert_eq!(
-            fields[1]
-                .metadata()
-                .get(EXPRESSION_META_KEY)
-                .map(String::as_str),
-            Some("first")
-        );
-    }
-
-    #[test]
-    fn test_blob_expression_transformation_does_not_inherit_semantics() {
-        let schema = Arc::new(ArrowSchema::new(vec![crate::blob("image", true)]));
-        let fields = plan(
-            schema,
-            &[("payload".to_string(), "coalesce(image, image)".to_string())],
-        )
-        .unwrap();
-
-        assert!(!fields[0].is_blob_v2());
-        assert_eq!(fields[0].data_type(), &DataType::LargeBinary);
-    }
-
    /// The binding reaches the schema only if `AllNulls` carries per-field
    /// metadata through the commit. The whole representation rests on it.
    #[tokio::test]
@@ -1864,40 +1582,6 @@ mod tests {
        assert!(declared(&table).await.is_empty());
    }

-    /// A batch may build on itself: one commit, and the later entry's inputs
-    /// name the earlier one.
-    #[tokio::test]
-    async fn test_a_declaration_may_read_one_declared_before_it() {
-        let table = table_with_ints("chain").await;
-        let before = table.version().await.unwrap();
-        add_computed(
-            &table,
-            &[("a".into(), "x + 1".into()), ("b".into(), "a * 2".into())],
-        )
-        .await
-        .unwrap();
-        assert_eq!(table.version().await.unwrap(), before + 1);
-        let declared = declared(&table).await;
-        assert_eq!(declared[1].name, "b");
-        assert_eq!(declared[1].inputs, vec!["a".to_string()]);
-
-        // Order is the dependency order; reading ahead is still unknown.
-        let err = add_computed(
-            &table,
-            &[("c".into(), "d + 1".into()), ("d".into(), "x + 1".into())],
-        )
-        .await
-        .unwrap_err();
-        assert!(matches!(err, Error::InvalidExpression { column, .. } if column == "c"));
-        assert!(
-            validate_declarations(
-                table.schema().await.unwrap(),
-                &[("e".into(), "random()".into())]
-            )
-            .is_err()
-        );
-    }
-
    /// A column added by an ordinary transform is materialized, not bound, so
    /// it carries no declaration to report.
    #[tokio::test]
@@ -28,9 +28,8 @@ pub(super) type PreparedIndex = (String, Box<dyn lance::index::IndexParams>, Ind
 use crate::index::Index;
 use crate::index::vector::{VectorIndex, suggested_num_sub_vectors};
 use crate::utils::{
-    resolve_lance_fts_field_path, supported_bitmap_data_type, supported_btree_data_type,
-    supported_fm_data_type, supported_fts_data_type, supported_label_list_data_type,
-    supported_vector_data_type,
+    supported_bitmap_data_type, supported_btree_data_type, supported_fm_data_type,
+    supported_fts_data_type, supported_label_list_data_type, supported_vector_data_type,
 };

 use super::NativeTable;
@@ -123,20 +122,7 @@ impl NativeTable {
        }
        self.dataset.ensure_mutable()?;
        let dataset = self.dataset.get().await?;
-        let (column, field) = if let Index::FTS(params) = &opts.index {
-            let resolved = resolve_lance_fts_field_path(dataset.schema(), &opts.columns[0])?;
-            if params.get_document_granularity().is_list_element() && resolved.list_depth == 0 {
-                return Err(Error::InvalidInput {
-                    message: format!(
-                        "FTS field path '{}' has no List layer and cannot use ListElement document granularity",
-                        resolved.canonical_path
-                    ),
-                });
-            }
-            (resolved.canonical_path, resolved.field)
-        } else {
-            Self::resolve_index_field(dataset.schema(), &opts.columns[0])?
-        };
+        let (column, field) = Self::resolve_index_field(dataset.schema(), &opts.columns[0])?;
        let params = self.make_index_params(&field, opts.index.clone()).await?;
        let index_type = self.get_index_type_for_field(&field, &opts.index);
        Ok((column, params, index_type))
@@ -450,7 +436,7 @@ mod tests {
    use crate::connection::ConnectBuilder;
    use crate::index::Index;
    use crate::index::scalar::{
-        BTreeIndexBuilder, BitmapIndexBuilder, DocumentGranularity, FmIndexBuilder, FtsIndexBuilder,
+        BTreeIndexBuilder, BitmapIndexBuilder, FmIndexBuilder, FtsIndexBuilder,
    };
    use crate::index::vector::{
        IvfHnswFlatIndexBuilder, IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder,
@@ -567,38 +553,6 @@ mod tests {
        job.cancel().await.unwrap();
    }

-    #[tokio::test]
-    async fn test_execute_async_validates_fts_input_before_starting_job() {
-        let conn = connect("memory://").execute().await.unwrap();
-        let batch =
-            record_batch!(("id", Int32, [1, 2]), ("text", Utf8, ["alpha", "beta"])).unwrap();
-        let table = conn.create_table("t", batch).execute().await.unwrap();
-
-        let missing = table
-            .create_index(&["missing"], Index::FTS(FtsIndexBuilder::default()))
-            .execute_async()
-            .await;
-        assert!(missing.is_err());
-
-        let invalid_type = table
-            .create_index(&["id"], Index::FTS(FtsIndexBuilder::default()))
-            .execute_async()
-            .await;
-        assert!(invalid_type.is_err());
-
-        let invalid_granularity = table
-            .create_index(
-                &["text"],
-                Index::FTS(
-                    FtsIndexBuilder::default()
-                        .document_granularity(DocumentGranularity::ListElement),
-                ),
-            )
-            .execute_async()
-            .await;
-        assert!(invalid_granularity.is_err());
-    }
-
    /// Concurrent waiters, and a wait issued after the job settled, all
    /// succeed once the build does.
    #[tokio::test]
@@ -7,7 +7,7 @@

 use std::sync::Arc;

-use arrow_schema::{DataType, Field, FieldRef, Fields};
+use arrow_schema::{DataType, Field, FieldRef};
 use datafusion::functions::core::{get_field, named_struct};
 use datafusion_common::ScalarValue;
 use datafusion_common::config::ConfigOptions;
@@ -35,17 +35,8 @@ pub(super) fn coerce_blob_expr(
        });
    };

-    let input_shape = match input_field.data_type() {
-        DataType::Null => {
-            let expr: Arc<dyn PhysicalExpr> = Arc::new(CastExpr::new(
-                input_expr,
-                table_field.data_type().clone(),
-                None,
-            ));
-            return Ok((expr, table_field.clone()));
-        }
-        DataType::Binary | DataType::LargeBinary | DataType::BinaryView => BlobInputShape::Bytes,
-        DataType::Utf8 | DataType::LargeUtf8 | DataType::Utf8View => BlobInputShape::String,
+    let input_struct_children = match input_field.data_type() {
+        DataType::Binary | DataType::LargeBinary | DataType::BinaryView => None,
        DataType::Struct(children) => {
            if !children
                .iter()
@@ -58,15 +49,13 @@ pub(super) fn coerce_blob_expr(
                    ),
                });
            }
-            BlobInputShape::Struct(children)
+            Some(children)
        }
        other => {
            return Err(Error::InvalidInput {
                message: format!(
                    "cannot coerce column '{}' with type {} into a blob v2 struct. \
-                     expected binary bytes (Binary, LargeBinary, BinaryView), \
-                     strings (Utf8, LargeUtf8, Utf8View), \
-                     or a Struct with a 'data' or 'uri' child",
+                     expected Binary, LargeBinary, BinaryView, or a Struct with a 'data' or 'uri' child",
                    table_field.name(),
                    other,
                ),
@@ -80,8 +69,9 @@ pub(super) fn coerce_blob_expr(
            declared.name().as_str(),
        ))));

-        let value: Arc<dyn PhysicalExpr> = match &input_shape {
-            BlobInputShape::Bytes => {
+        let value: Arc<dyn PhysicalExpr> = match input_struct_children {
+            // Raw binary lands in `data` and everything else is a typed null.
+            None => {
                if declared.name() == "data" {
                    Arc::new(CastExpr::new(
                        input_expr.clone(),
@@ -92,43 +82,30 @@ pub(super) fn coerce_blob_expr(
                    typed_null(declared.data_type())?
                }
            }
-            BlobInputShape::String => {
-                if declared.name() == "uri" {
-                    Arc::new(CastExpr::new(
-                        input_expr.clone(),
-                        declared.data_type().clone(),
-                        None,
-                    ))
-                } else {
-                    typed_null(declared.data_type())?
-                }
-            }
-            BlobInputShape::Struct(children) => {
-                match children.iter().find(|c| c.name() == declared.name()) {
-                    Some(child) => {
-                        let field_expr: Arc<dyn PhysicalExpr> = Arc::new(ScalarFunctionExpr::new(
-                            &format!("get_field({})", declared.name()),
-                            get_field(),
-                            vec![
-                                input_expr.clone(),
-                                Arc::new(Literal::new(ScalarValue::from(declared.name().as_str()))),
-                            ],
-                            Arc::new(child.as_ref().clone()),
-                            config.clone(),
-                        ));
-                        if child.data_type() == declared.data_type() {
-                            field_expr
-                        } else {
-                            Arc::new(CastExpr::new(
-                                field_expr,
-                                declared.data_type().clone(),
-                                None,
-                            ))
-                        }
+            Some(children) => match children.iter().find(|c| c.name() == declared.name()) {
+                Some(child) => {
+                    let field_expr: Arc<dyn PhysicalExpr> = Arc::new(ScalarFunctionExpr::new(
+                        &format!("get_field({})", declared.name()),
+                        get_field(),
+                        vec![
+                            input_expr.clone(),
+                            Arc::new(Literal::new(ScalarValue::from(declared.name().as_str()))),
+                        ],
+                        Arc::new(child.as_ref().clone()),
+                        config.clone(),
+                    ));
+                    if child.data_type() == declared.data_type() {
+                        field_expr
+                    } else {
+                        Arc::new(CastExpr::new(
+                            field_expr,
+                            declared.data_type().clone(),
+                            None,
+                        ))
                    }
-                    None => typed_null(declared.data_type())?,
                }
-            }
+                None => typed_null(declared.data_type())?,
+            },
        };
        ns_args.push(value);
    }
@@ -143,12 +120,6 @@ pub(super) fn coerce_blob_expr(
    Ok((expr, table_field.clone()))
 }

-enum BlobInputShape<'a> {
-    Bytes,
-    String,
-    Struct(&'a Fields),
-}
-
 fn typed_null(data_type: &DataType) -> Result<Arc<dyn PhysicalExpr>> {
    let scalar = ScalarValue::try_from(data_type).map_err(|e| Error::InvalidInput {
        message: format!("cannot build null literal for blob child type {data_type}: {e}"),
@@ -163,7 +134,7 @@ mod tests {
    use crate::blob::blob;
    use arrow_array::{
        Array, ArrayRef, BinaryArray, BinaryViewArray, Int32Array, Int64Array, LargeBinaryArray,
-        NullArray, RecordBatch, StringArray, StringViewArray, StructArray, UInt8Array, UInt64Array,
+        RecordBatch, StringArray, StructArray, UInt8Array, UInt64Array,
    };
    use arrow_schema::Schema;
    use datafusion::prelude::SessionContext;
@@ -287,18 +258,6 @@ mod tests {
        assert_eq!(data.value(0), b"view");
    }

-    #[tokio::test]
-    async fn null_column_coerces_to_all_null_blob_struct() {
-        let batch = batch_with_image(
-            Field::new("image", DataType::Null, true),
-            Arc::new(NullArray::new(2)),
-        );
-        let coerced = coerce(batch, &blob_table_schema()).await;
-        let image = image_struct(&coerced);
-        assert!(image.is_null(0));
-        assert!(image.is_null(1));
-    }
-
    #[tokio::test]
    async fn binary_nulls_stay_null_after_coercion() {
        let batch = batch_with_image(
@@ -477,78 +436,14 @@ mod tests {
    #[tokio::test]
    async fn unsupported_input_type_is_rejected_with_column_name() {
        let batch = batch_with_image(
-            Field::new("image", DataType::Int64, true),
-            Arc::new(Int64Array::from(vec![42])),
+            Field::new("image", DataType::Utf8, true),
+            Arc::new(StringArray::from(vec!["not bytes"])),
        );
        let err = coerce_err(batch, &blob_table_schema()).await;
        assert!(matches!(err, Error::InvalidInput { .. }), "got {err:?}");
        assert!(err.to_string().contains("image"));
    }

-    #[tokio::test]
-    async fn utf8_string_coerces_to_uri_child() {
-        let batch = batch_with_image(
-            Field::new("image", DataType::Utf8, true),
-            Arc::new(StringArray::from(vec![Some("s3://bucket/key"), None])),
-        );
-        let coerced = coerce(batch, &blob_table_schema()).await;
-        let image = image_struct(&coerced);
-        let uri: &StringArray = image
-            .column_by_name("uri")
-            .unwrap()
-            .as_any()
-            .downcast_ref()
-            .unwrap();
-        assert_eq!(uri.value(0), "s3://bucket/key");
-        assert!(image.column_by_name("data").unwrap().is_null(0));
-        assert!(uri.is_null(1));
-    }
-
-    #[tokio::test]
-    async fn large_utf8_string_coerces_into_four_child_blob_layout() {
-        use arrow_array::LargeStringArray;
-
-        let table_schema = Schema::new(vec![
-            Field::new("id", DataType::Int64, false),
-            wide_blob_field("image"),
-        ]);
-        let batch = batch_with_image(
-            Field::new("image", DataType::LargeUtf8, true),
-            Arc::new(LargeStringArray::from(vec!["file:///tmp/blob.bin"])),
-        );
-        let coerced = coerce(batch, &table_schema).await;
-        let image = image_struct(&coerced);
-        assert_eq!(image.num_columns(), 4);
-        let uri: &StringArray = image
-            .column_by_name("uri")
-            .unwrap()
-            .as_any()
-            .downcast_ref()
-            .unwrap();
-        assert_eq!(uri.value(0), "file:///tmp/blob.bin");
-        assert!(image.column_by_name("data").unwrap().is_null(0));
-        assert!(image.column_by_name("position").unwrap().is_null(0));
-        assert!(image.column_by_name("size").unwrap().is_null(0));
-    }
-
-    #[tokio::test]
-    async fn utf8_view_string_coerces_to_uri_child() {
-        let batch = batch_with_image(
-            Field::new("image", DataType::Utf8View, true),
-            Arc::new(StringViewArray::from(vec![Some("s3://bucket/view-key")])),
-        );
-        let coerced = coerce(batch, &blob_table_schema()).await;
-        let image = image_struct(&coerced);
-        let uri: &StringArray = image
-            .column_by_name("uri")
-            .unwrap()
-            .as_any()
-            .downcast_ref()
-            .unwrap();
-        assert_eq!(uri.value(0), "s3://bucket/view-key");
-        assert!(image.column_by_name("data").unwrap().is_null(0));
-    }
-
    #[tokio::test]
    async fn blob_metadata_survives_cast_of_sibling_column() {
        let batch = RecordBatch::try_new(
@@ -32,10 +32,6 @@ struct DatasetState {
    /// `Some(version)` = pinned to a specific version (time travel),
    /// `None` = tracking latest.
    pinned_version: Option<u64>,
-    /// Whether the pin is an internal query snapshot rather than user-visible
-    /// time travel. Query snapshots remain read-only but preserve MemWAL read
-    /// semantics.
-    query_snapshot: bool,
 }

 #[derive(Debug, Clone)]
@@ -74,7 +70,6 @@ impl DatasetConsistencyWrapper {
            state: Arc::new(Mutex::new(DatasetState {
                dataset,
                pinned_version: None,
-                query_snapshot: false,
            })),
            consistency,
            shard_writer: Arc::new(ShardWriterCache::default()),
@@ -98,36 +93,6 @@ impl DatasetConsistencyWrapper {
        wrapper
    }

-    /// Create an independent read-only wrapper pinned to the current dataset
-    /// while retaining this wrapper's live MemWAL read context.
-    pub async fn new_query_snapshot(&self) -> Result<Self> {
-        // Apply the configured consistency policy before taking the snapshot.
-        // The returned dataset is intentionally discarded: a checkout may race
-        // after this await, so the dataset and its pin provenance must instead
-        // be cloned together from one authoritative state sample below.
-        self.get().await?;
-
-        let (dataset, query_snapshot) = {
-            let state = self.state.lock()?;
-            // Preserve user time travel so the MemWAL safety guard still sees
-            // it. Latest and already-internal snapshots remain internal pins.
-            (
-                state.dataset.clone(),
-                state.query_snapshot || state.pinned_version.is_none(),
-            )
-        };
-        let version = dataset.version().version;
-        Ok(Self {
-            state: Arc::new(Mutex::new(DatasetState {
-                dataset,
-                pinned_version: Some(version),
-                query_snapshot,
-            })),
-            consistency: ConsistencyMode::Lazy,
-            shard_writer: self.shard_writer.clone(),
-        })
-    }
-
    /// The MemWAL `ShardWriter` cache co-located with this dataset.
    pub(crate) fn shard_writer(&self) -> &Arc<ShardWriterCache> {
        &self.shard_writer
@@ -204,7 +169,6 @@ impl DatasetConsistencyWrapper {
        let mut state = self.state.lock()?;
        state.dataset = Arc::new(new_dataset);
        state.pinned_version = None;
-        state.query_snapshot = false;
        drop(state);
        if let ConsistencyMode::Eventual(bg_cache) = &self.consistency {
            bg_cache.invalidate();
@@ -238,10 +202,10 @@ impl DatasetConsistencyWrapper {

    /// Returns the version, if in time travel mode, or None otherwise.
    pub fn time_travel_version(&self) -> Option<u64> {
-        let state = self.state.lock().unwrap_or_else(|e| e.into_inner());
-        (!state.query_snapshot)
-            .then_some(state.pinned_version)
-            .flatten()
+        self.state
+            .lock()
+            .unwrap_or_else(|e| e.into_inner())
+            .pinned_version
    }

    /// Convert into a wrapper in latest version mode.
@@ -261,7 +225,6 @@ impl DatasetConsistencyWrapper {
        if state.pinned_version.is_some() {
            state.dataset = Arc::new(new_dataset);
            state.pinned_version = None;
-            state.query_snapshot = false;
        }
        drop(state);
        if let ConsistencyMode::Eventual(bg_cache) = &self.consistency {
@@ -297,7 +260,6 @@ impl DatasetConsistencyWrapper {
        let mut state = self.state.lock()?;
        state.dataset = Arc::new(new_dataset);
        state.pinned_version = Some(version_value);
-        state.query_snapshot = false;
        Ok(())
    }

@@ -499,29 +461,6 @@ mod tests {
        assert_eq!(wrapper.time_travel_version(), Some(1));
    }

-    #[tokio::test]
-    async fn test_query_snapshot_samples_dataset_and_pin_together() {
-        let dir = tempfile::tempdir().unwrap();
-        let uri = dir.path().to_str().unwrap();
-        let ds = create_test_dataset(uri).await;
-
-        let wrapper = DatasetConsistencyWrapper::new_latest(ds, None);
-        wrapper.as_time_travel(1u64).await.unwrap();
-        let stale_time_travel_dataset = wrapper.get().await.unwrap();
-
-        append_to_dataset(uri).await;
-        wrapper.as_latest().await.unwrap();
-
-        let snapshot = wrapper.new_query_snapshot().await.unwrap();
-        let snapshot_dataset = snapshot.get().await.unwrap();
-        assert_eq!(snapshot_dataset.version().version, 2);
-        assert_ne!(
-            snapshot_dataset.version().version,
-            stale_time_travel_dataset.version().version
-        );
-        assert_eq!(snapshot.time_travel_version(), None);
-    }
-
    #[tokio::test]
    async fn test_as_latest_from_time_travel() {
        let dir = tempfile::tempdir().unwrap();
@@ -31,9 +31,8 @@ pub(crate) async fn execute_delete(
    table.dataset.ensure_mutable()?;
    match predicate {
        Predicate::String(s) => {
-            let predicate = crate::expr::canonicalize_sql_predicate(s)?;
            let mut dataset = (*table.dataset.get().await?).clone();
-            let delete_result = dataset.delete(&predicate).boxed().await?;
+            let delete_result = dataset.delete(s).boxed().await?;
            let num_deleted_rows = delete_result.num_deleted_rows;
            let version = dataset.version().version;
            table.dataset.update(dataset);
@@ -220,32 +220,9 @@ impl MergeInsertBuilder {
    ///
    /// Returns version and statistics about the merge operation including the number of rows
    /// inserted, updated, and deleted.
-    pub async fn execute(
-        mut self,
-        new_data: Box<dyn RecordBatchReader + Send>,
-    ) -> Result<MergeResult> {
-        self.canonicalize_filters()?;
+    pub async fn execute(self, new_data: Box<dyn RecordBatchReader + Send>) -> Result<MergeResult> {
        self.table.clone().merge_insert(self, new_data).await
    }
-
-    pub(crate) fn canonicalize_filters(&mut self) -> Result<()> {
-        self.when_matched_update_all_filt =
-            canonicalize_merge_filter(self.when_matched_update_all_filt.take())?;
-        self.when_not_matched_by_source_delete_filt =
-            canonicalize_merge_filter(self.when_not_matched_by_source_delete_filt.take())?;
-        Ok(())
-    }
-}
-
-fn canonicalize_merge_filter(filter: Option<MergeFilter>) -> Result<Option<MergeFilter>> {
-    filter
-        .map(|filter| match filter {
-            MergeFilter::Sql(predicate) => {
-                crate::expr::canonicalize_sql_predicate(&predicate).map(MergeFilter::Sql)
-            }
-            filter @ MergeFilter::Expr(_) => Ok(filter),
-        })
-        .transpose()
 }

 /// Internal implementation of the merge insert logic
@@ -253,10 +230,9 @@ fn canonicalize_merge_filter(filter: Option<MergeFilter>) -> Result<Option<Merge
 /// This logic was moved from NativeTable::merge_insert to keep table.rs clean.
 pub(crate) async fn execute_merge_insert(
    table: &NativeTable,
-    mut params: MergeInsertBuilder,
+    params: MergeInsertBuilder,
    new_data: Box<dyn RecordBatchReader + Send>,
 ) -> Result<MergeResult> {
-    params.canonicalize_filters()?;
    super::computed_columns::ensure_no_function_bindings_for_mutation(
        table.schema().await?.as_ref(),
        "merge_insert",
@@ -1080,44 +1056,6 @@ mod lsm_tests {
        );
    }

-    #[tokio::test]
-    async fn query_snapshot_preserves_lsm_read_semantics() {
-        let dir = tempdir().unwrap();
-        let table = id_value_table(&dir).await;
-        table
-            .set_lsm_write_spec(LsmWriteSpec::unsharded())
-            .await
-            .unwrap();
-        lsm_upsert(&table, vec![4, 5]).await;
-
-        let snapshot = table.query_snapshot().await.unwrap();
-        let rows = collect_id_value(snapshot.query().execute().await.unwrap()).await;
-        assert_eq!(
-            rows.iter().map(|(id, _)| *id).collect::<Vec<_>>(),
-            vec![1, 2, 3, 4, 5]
-        );
-    }
-
-    #[tokio::test]
-    async fn query_snapshot_preserves_time_travel_lsm_guard() {
-        let dir = tempdir().unwrap();
-        let table = id_value_table(&dir).await;
-        table
-            .set_lsm_write_spec(LsmWriteSpec::unsharded())
-            .await
-            .unwrap();
-        lsm_upsert(&table, vec![4]).await;
-
-        let version = table.version().await.unwrap();
-        table.checkout(version).await.unwrap();
-        let direct_error = table.query().execute().await.err().unwrap();
-        assert!(matches!(direct_error, Error::NotSupported { .. }));
-
-        let snapshot = table.query_snapshot().await.unwrap();
-        let snapshot_error = snapshot.query().execute().await.err().unwrap();
-        assert!(matches!(snapshot_error, Error::NotSupported { .. }));
-    }
-
    #[tokio::test]
    async fn lsm_read_dedup_newest_wins() {
        let dir = tempdir().unwrap();
@@ -15,7 +15,7 @@ use lance_index::optimize::OptimizeOptions;
 use log::info;

 pub use chrono::Duration;
-pub use lance::dataset::optimize::{CompactionMode, CompactionOptions, IndexRemapMode};
+pub use lance::dataset::optimize::CompactionOptions;

 use super::NativeTable;
 use crate::error::Result;
@@ -17,11 +17,11 @@ use arrow::array::{AsArray, FixedSizeListBuilder, Float32Builder};
 use arrow::datatypes::{Float32Type, UInt8Type};
 use arrow_array::Array;
 use arrow_schema::{DataType, Schema};
-use datafusion_common::{Column, DataFusionError, SchemaError};
 use datafusion_physical_plan::ExecutionPlan;
 use datafusion_physical_plan::projection::ProjectionExec;
 use datafusion_physical_plan::repartition::RepartitionExec;
 use datafusion_physical_plan::union::UnionExec;
+use futures::future::try_join_all;
 use lance::dataset::mem_wal::DatasetMemWalExt;
 use lance::dataset::scanner::DatasetRecordBatchStream;
 use lance::dataset::scanner::Scanner;
@@ -45,22 +45,6 @@ impl AnyQuery {
            Self::VectorQuery(query) => &query.base,
        }
    }
-
-    fn base_mut(&mut self) -> &mut QueryRequest {
-        match self {
-            Self::Query(query) => query,
-            Self::VectorQuery(query) => &mut query.base,
-        }
-    }
-
-    /// Canonicalize any raw SQL filter immediately before backend dispatch.
-    pub(crate) fn canonicalized(&self) -> Result<Self> {
-        let mut query = self.clone();
-        if let Some(QueryFilter::Sql(predicate)) = &mut query.base_mut().filter {
-            *predicate = crate::expr::canonicalize_sql_predicate(predicate)?;
-        }
-        Ok(query)
-    }
 }

 //Decide between namespace or local
@@ -69,16 +53,15 @@ pub async fn execute_query(
    query: &AnyQuery,
    options: QueryExecutionOptions,
 ) -> Result<DatasetRecordBatchStream> {
-    let query = query.canonicalized()?;
    // QueryTable pushdown runs the query server-side, but only on the main
    // branch: the namespace request carries no branch yet, so a branch handle
    // must fall through to local execution.
-    if can_execute_namespace_query(table, &query).await?
+    if can_execute_namespace_query(table, query).await?
        && let Some(ref namespace_client) = table.namespace_client
    {
-        return execute_namespace_query(table, namespace_client.clone(), &query, options).await;
+        return execute_namespace_query(table, namespace_client.clone(), query, options).await;
    }
-    execute_generic_query(table, &query, options).await
+    execute_generic_query(table, query, options).await
 }

 async fn can_execute_namespace_query(table: &NativeTable, query: &AnyQuery) -> Result<bool> {
@@ -153,10 +136,9 @@ pub async fn create_plan(
    query: &AnyQuery,
    options: QueryExecutionOptions,
 ) -> Result<Arc<dyn ExecutionPlan>> {
-    let query = query.canonicalized()?;
    let query = match query {
-        AnyQuery::VectorQuery(query) => query,
-        AnyQuery::Query(query) => VectorQueryRequest::from_plain_query(query),
+        AnyQuery::VectorQuery(query) => query.clone(),
+        AnyQuery::Query(query) => VectorQueryRequest::from_plain_query(query.clone()),
    };
    query.base.check_filter()?;

@@ -188,48 +170,26 @@ pub async fn create_plan(
    let mut column = query.column.clone();

    let mut query_vector = query.query_vector.first().cloned();
-    let mut is_batch_query = false;
    if query.query_vector.len() > 1 {
        if column.is_none() {
            // Infer a vector column with the same dimension of the query vector.
-            let arrow_schema = Schema::from(schema);
+            let arrow_schema = Schema::from(ds_ref.schema());
            column = Some(default_vector_column(
                &arrow_schema,
                Some(query.query_vector[0].len() as i32),
            )?);
        }
        let vector_field = schema.field(column.as_ref().unwrap()).unwrap();
-        let (_, element_type) =
-            lance::index::vector::utils::get_vector_type(schema, column.as_ref().unwrap())?;
-        let is_binary = matches!(element_type, DataType::UInt8);
-        if matches!(vector_field.data_type(), DataType::List(_))
-            || (query.base.offset.unwrap_or(0) == 0 && !is_binary)
-        {
-            // Lance distinguishes these cases from the vector column type: a
-            // list-like query against a List column is one multivector query,
-            // while the same query against a FixedSizeList column is a batch of
-            // independent queries. The batch path shares a single flat scan and
-            // bounds retained candidate data instead of running one scan per
-            // query vector.
+        if let DataType::List(_) = vector_field.data_type() {
+            // Multivector handling: concatenate into FixedSizeList<FixedSizeList<_>>
            let vectors = query
                .query_vector
                .iter()
                .map(|arr| arr.as_ref())
                .collect::<Vec<_>>();
            let dim = vectors[0].len();
-            if let Some((query_index, actual_dim)) = vectors
-                .iter()
-                .enumerate()
-                .find_map(|(index, vector)| (vector.len() != dim).then_some((index, vector.len())))
-            {
-                return Err(Error::InvalidInput {
-                    message: format!(
-                        "query vector at index {query_index} has dimension {actual_dim}, expected {dim}"
-                    ),
-                });
-            }
            let mut fsl_builder = FixedSizeListBuilder::with_capacity(
-                Float32Builder::with_capacity(dim * vectors.len()),
+                Float32Builder::with_capacity(dim),
                dim as i32,
                vectors.len(),
            );
@@ -240,12 +200,8 @@ pub async fn create_plan(
                fsl_builder.append(true);
            }
            query_vector = Some(Arc::new(fsl_builder.finish()));
-            is_batch_query = !matches!(vector_field.data_type(), DataType::List(_));
        } else {
-            // Lance's batch path has no per-query offset, and its binary path
-            // requires primitive UInt8 queries rather than a fixed-size list.
-            // Keep the prior plan shape for these cases so offsets are applied
-            // per query and binary query vectors retain their primitive shape.
+            // Multiple query vectors: create a plan for each and union them
            let query_vecs = query.query_vector.clone();
            let plan_futures = query_vecs
                .into_iter()
@@ -258,7 +214,7 @@ pub async fn create_plan(
                    }
                })
                .collect::<Vec<_>>();
-            let plans = futures::future::try_join_all(plan_futures).await?;
+            let plans = try_join_all(plan_futures).await?;
            return create_multi_vector_plan(plans);
        }
    }
@@ -269,7 +225,7 @@ pub async fn create_plan(
        let column = if let Some(col) = column {
            col
        } else {
-            let arrow_schema = Schema::from(schema);
+            let arrow_schema = Schema::from(ds_ref.schema());
            default_vector_column(&arrow_schema, Some(query_vector.len() as i32))?
        };

@@ -295,14 +251,10 @@ pub async fn create_plan(
        }
    }

-    // For a batch query, `nearest` already applies k to each query vector.
-    // Adding Scanner's global limit would truncate the combined result to k rows.
-    if !is_batch_query {
-        scanner.limit(
-            query.base.limit.map(|limit| limit as i64),
-            query.base.offset.map(|offset| offset as i64),
-        )?;
-    }
+    scanner.limit(
+        query.base.limit.map(|limit| limit as i64),
+        query.base.offset.map(|offset| offset as i64),
+    )?;

    if let Some(ef) = query.ef {
        scanner.ef(ef);
@@ -375,97 +327,7 @@ pub async fn create_plan(
        scanner.order_by(Some(order_by.clone()))?;
    }

-    scanner
-        .create_plan()
-        .await
-        .map_err(|error| enrich_lance_field_not_found(error, schema))
-}
-
-/// Replace DataFusion's top-level field candidates with qualified leaf paths.
-///
-/// DataFusion resolves nested fields but its `FieldNotFound` error only lists the
-/// top-level Arrow fields. This makes a missing leaf look unavailable even when it
-/// exists below a struct. Keep every other Lance/DataFusion error unchanged and
-/// enrich only this one schema error at the LanceDB query boundary.
-fn enrich_lance_field_not_found(
-    error: lance::Error,
-    schema: &lance_core::datatypes::Schema,
-) -> Error {
-    let Some(field) = find_missing_field(&error) else {
-        return error.into();
-    };
-    field_not_found_error(field, &Schema::from(schema))
-}
-
-fn field_not_found_diagnostic(
-    error: &(dyn std::error::Error + 'static),
-    schema: &Schema,
-) -> Option<Error> {
-    let field = find_missing_field(error)?;
-    Some(field_not_found_error(field, schema))
-}
-
-fn field_not_found_error(field: &Column, schema: &Schema) -> Error {
-    let valid_fields = leaf_field_paths(schema);
-    let mut message = format!("Schema error: No field named {}", field.quoted_flat_name());
-    if !valid_fields.is_empty() {
-        message.push_str(". Valid fields are ");
-        message.push_str(&valid_fields.join(", "));
-    }
-    message.push('.');
-
-    Error::InvalidInput { message }
-}
-
-fn find_missing_field<'a>(error: &'a (dyn std::error::Error + 'static)) -> Option<&'a Column> {
-    if let Some(DataFusionError::SchemaError(schema_error, _)) =
-        error.downcast_ref::<DataFusionError>()
-        && let SchemaError::FieldNotFound { field, .. } = schema_error.as_ref()
-    {
-        return Some(field);
-    }
-
-    error.source().and_then(find_missing_field)
-}
-
-fn leaf_field_paths(schema: &Schema) -> Vec<String> {
-    fn format_segment(segment: &str) -> String {
-        // Quote every segment instead of maintaining a SQL keyword list. Bare
-        // lowercase names such as `true` can be parsed as expressions rather
-        // than identifiers, while backticks preserve all field names in both
-        // local SQL parsers.
-        format!("`{}`", segment.replace('`', "``"))
-    }
-
-    fn visit(fields: &arrow_schema::Fields, path: &mut Vec<String>, paths: &mut Vec<String>) {
-        for field in fields {
-            // Neither local planner can address an empty field-path segment,
-            // even when it is backtick-quoted. Do not advertise leaves beneath
-            // such a segment as valid filter fields.
-            if field.name().is_empty() {
-                continue;
-            }
-            path.push(field.name().clone());
-            match field.data_type() {
-                DataType::Struct(children) if !children.is_empty() => {
-                    visit(children, path, paths);
-                }
-                _ => {
-                    paths.push(
-                        path.iter()
-                            .map(|segment| format_segment(segment))
-                            .collect::<Vec<_>>()
-                            .join("."),
-                    );
-                }
-            }
-            path.pop();
-        }
-    }
-
-    let mut paths = Vec::new();
-    visit(schema.fields(), &mut Vec::new(), &mut paths);
-    paths
+    Ok(scanner.create_plan().await?)
 }

 //Helper functions below
@@ -825,10 +687,7 @@ async fn parse_arrow_ipc_response(bytes: bytes::Bytes) -> Result<DatasetRecordBa
 #[cfg(test)]
 #[allow(deprecated)]
 mod tests {
-    use arrow_array::{
-        ArrayRef, FixedSizeListArray, Float32Array, Int32Array, RecordBatch, StringArray,
-        StructArray,
-    };
+    use arrow_array::{ArrayRef, FixedSizeListArray, Float32Array};
    use futures::TryStreamExt;
    use lance_arrow::FixedSizeListArrayExt;
    use std::sync::{
@@ -837,8 +696,7 @@ mod tests {
    };

    use super::*;
-    use crate::query::{ExecutableQuery, QueryBase, QueryExecutionOptions, QueryRequest};
-    use crate::table::BaseTable;
+    use crate::query::{QueryExecutionOptions, QueryRequest};

    fn fixed_size_list_array(values: Vec<f32>, dimension: i32) -> FixedSizeListArray {
        FixedSizeListArray::try_new_from_values(Float32Array::from(values), dimension).unwrap()
@@ -978,6 +836,7 @@ mod tests {
    async fn test_execute_query_local_routing() {
        use crate::connect;
        use crate::table::query::execute_query;
+        use arrow_array::{Int32Array, RecordBatch};
        use arrow_schema::{DataType, Field, Schema};

        let conn = connect("memory://").execute().await.unwrap();
@@ -1017,164 +876,6 @@ mod tests {
        assert_eq!(count, 2); // 4 and 5
    }

-    #[tokio::test]
-    async fn test_missing_filter_field_lists_nested_fields_in_local_planners() {
-        use crate::connect;
-        use arrow_schema::{DataType, Field, Schema};
-
-        let conn = connect("memory://").execute().await.unwrap();
-        let metadata = Arc::new(StructArray::from(vec![
-            (
-                Arc::new(Field::new("year", DataType::Int32, false)),
-                Arc::new(Int32Array::from(vec![2024])) as ArrayRef,
-            ),
-            (
-                Arc::new(Field::new("genre", DataType::Utf8, false)),
-                Arc::new(StringArray::from(vec!["fiction"])) as ArrayRef,
-            ),
-            (
-                Arc::new(Field::new("Title", DataType::Int32, false)),
-                Arc::new(Int32Array::from(vec![7])) as ArrayRef,
-            ),
-            (
-                Arc::new(Field::new("true", DataType::Int32, false)),
-                Arc::new(Int32Array::from(vec![8])) as ArrayRef,
-            ),
-            (
-                Arc::new(Field::new("", DataType::Int32, false)),
-                Arc::new(Int32Array::from(vec![10])) as ArrayRef,
-            ),
-        ]));
-        let vector = Arc::new(fixed_size_list_array(vec![0.0, 1.0], 2));
-        let schema = Arc::new(Schema::new(vec![
-            Field::new("id", DataType::Int32, false),
-            Field::new("vector", vector.data_type().clone(), false),
-            Field::new("content", DataType::Utf8, false),
-            Field::new("metadata", metadata.data_type().clone(), false),
-        ]));
-        let batch = RecordBatch::try_new(
-            schema,
-            vec![
-                Arc::new(Int32Array::from(vec![1])),
-                vector,
-                Arc::new(StringArray::from(vec!["example"])),
-                metadata,
-            ],
-        )
-        .unwrap();
-        let table = conn
-            .create_table("nested_error", batch)
-            .execute()
-            .await
-            .unwrap();
-
-        let error = table
-            .query()
-            .only_if("year = 2024")
-            .execute()
-            .await
-            .err()
-            .expect("query should reject the unqualified nested field");
-        let case_sensitive_path = "`metadata`.`Title`";
-        let keyword_path = "`metadata`.`true`";
-        let expected = format!(
-            "No field named year. Valid fields are `id`, `vector`, `content`, `metadata`.`year`, `metadata`.`genre`, {case_sensitive_path}, {keyword_path}."
-        );
-
-        assert!(
-            error.to_string().contains(&expected),
-            "unexpected error: {error}"
-        );
-        for (path, value) in [(case_sensitive_path, 7), (keyword_path, 8)] {
-            table
-                .query()
-                .only_if(format!("{path} = {value}"))
-                .execute()
-                .await
-                .expect("the path advertised by the diagnostic should be reusable");
-        }
-
-        table.set_unenforced_primary_key(["id"]).await.unwrap();
-        table
-            .set_lsm_write_spec(crate::table::LsmWriteSpec::unsharded())
-            .await
-            .unwrap();
-        let lsm_error = table
-            .query()
-            .only_if("year = 2024")
-            .execute()
-            .await
-            .err()
-            .expect("LSM query should reject the unqualified nested field");
-
-        assert!(
-            lsm_error.to_string().contains(&expected),
-            "unexpected LSM error: {lsm_error}"
-        );
-        for (path, value) in [(case_sensitive_path, 7), (keyword_path, 8)] {
-            table
-                .query()
-                .only_if(format!("{path} = {value}"))
-                .execute()
-                .await
-                .expect("the path advertised by the diagnostic should be reusable in LSM queries");
-        }
-    }
-
-    #[test]
-    fn test_leaf_field_paths_preserve_arbitrary_depth() {
-        use arrow_schema::{DataType, Field, Schema};
-
-        fn nested_field(path: &[&str]) -> Field {
-            let mut segments = path.iter().rev();
-            let mut field = Field::new(
-                *segments.next().expect("path must have a leaf"),
-                DataType::Int32,
-                false,
-            );
-            for segment in segments {
-                field = Field::new(*segment, DataType::Struct(vec![field].into()), false);
-            }
-            field
-        }
-
-        let schema = Schema::new(vec![
-            nested_field(&["a", "b", "c", "d", "e"]),
-            nested_field(&["metadata", "child.with.dot"]),
-            nested_field(&["metadata", "Title"]),
-            nested_field(&["metadata", "123child"]),
-            nested_field(&["metadata", "child`tick"]),
-            nested_field(&["metadata", ""]),
-            nested_field(&["", "child"]),
-        ]);
-
-        assert_eq!(
-            leaf_field_paths(&schema),
-            vec![
-                "`a`.`b`.`c`.`d`.`e`",
-                "`metadata`.`child.with.dot`",
-                "`metadata`.`Title`",
-                "`metadata`.`123child`",
-                "`metadata`.`child``tick`",
-            ]
-        );
-
-        let source = DataFusionError::SchemaError(
-            Box::new(SchemaError::FieldNotFound {
-                field: Box::new(Column::from_name("missing")),
-                valid_fields: Vec::new(),
-            }),
-            Box::new(None),
-        );
-        let error = field_not_found_diagnostic(&source, &schema).unwrap();
-        assert!(
-            error.to_string().contains(
-                "Valid fields are `a`.`b`.`c`.`d`.`e`, `metadata`.`child.with.dot`, `metadata`.`Title`, `metadata`.`123child`, `metadata`.`child``tick`"
-            ),
-            "unexpected error: {error}"
-        );
-    }
-
    #[derive(Debug, Default)]
    struct CountingNamespaceClient {
        query_table_calls: AtomicUsize,
@@ -1188,56 +889,10 @@ mod tests {

        async fn query_table(&self, _request: NsQueryTableRequest) -> lance::Result<bytes::Bytes> {
            self.query_table_calls.fetch_add(1, Ordering::SeqCst);
-            panic!("query must not be pushed down to namespace query_table");
+            panic!("approx_mode queries must not be pushed down to namespace query_table");
        }
    }

-    #[tokio::test]
-    async fn test_execute_query_pinned_snapshot_with_namespace_pushdown_runs_locally() {
-        use crate::connect;
-        use arrow_array::{Int32Array, RecordBatch};
-        use arrow_schema::{DataType, Field, Schema};
-
-        let conn = connect("memory://").execute().await.unwrap();
-        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
-        let batch = RecordBatch::try_new(
-            schema,
-            vec![Arc::new(Int32Array::from(vec![1, 2, 3, 4, 5]))],
-        )
-        .unwrap();
-        let table = conn
-            .create_table("test_pinned_namespace_fallback", vec![batch])
-            .execute()
-            .await
-            .unwrap();
-
-        let namespace_client = Arc::new(CountingNamespaceClient::default());
-        let mut native_table = table.as_native().unwrap().clone();
-        native_table.namespace_client = Some(namespace_client.clone());
-        native_table
-            .pushdown_operations
-            .insert(NamespaceClientPushdownOperation::QueryTable);
-
-        let snapshot = native_table.checkout_current().await.unwrap();
-        let snapshot = snapshot.as_any().downcast_ref::<NativeTable>().unwrap();
-        assert!(snapshot.dataset.time_travel_version().is_some());
-
-        let query = AnyQuery::Query(QueryRequest {
-            filter: Some(QueryFilter::Sql("id > 3".to_string())),
-            ..Default::default()
-        });
-        let stream = execute_query(snapshot, &query, QueryExecutionOptions::default())
-            .await
-            .unwrap();
-        let batches = stream.try_collect::<Vec<_>>().await.unwrap();
-
-        assert_eq!(
-            batches.iter().map(|batch| batch.num_rows()).sum::<usize>(),
-            2
-        );
-        assert_eq!(namespace_client.query_table_calls.load(Ordering::SeqCst), 0);
-    }
-
    #[tokio::test]
    async fn test_execute_query_approx_mode_with_namespace_pushdown_runs_locally() {
        use crate::connect;
@@ -1355,38 +1010,7 @@ mod tests {
    }

    #[tokio::test]
-    async fn test_query_snapshot_disables_namespace_pushdown() {
-        use crate::connect;
-        use crate::table::BaseTable;
-        use arrow_array::{Int32Array, RecordBatch};
-        use arrow_schema::{DataType, Field, Schema};
-
-        let conn = connect("memory://").execute().await.unwrap();
-        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
-        let batch =
-            RecordBatch::try_new(schema, vec![Arc::new(Int32Array::from(vec![1, 2, 3]))]).unwrap();
-        let table = conn
-            .create_table("test_snapshot_namespace_fallback", vec![batch])
-            .execute()
-            .await
-            .unwrap();
-        let mut native_table = table.as_native().unwrap().clone();
-        native_table.namespace_client = Some(Arc::new(CountingNamespaceClient::default()));
-        native_table
-            .pushdown_operations
-            .insert(NamespaceClientPushdownOperation::QueryTable);
-
-        let snapshot = BaseTable::query_snapshot(&native_table).await.unwrap();
-        let snapshot = snapshot.as_any().downcast_ref::<NativeTable>().unwrap();
-        assert!(
-            !can_execute_namespace_query(snapshot, &AnyQuery::Query(QueryRequest::default()),)
-                .await
-                .unwrap()
-        );
-    }
-
-    #[tokio::test]
-    async fn test_create_plan_batch_vector_uses_shared_scan() {
+    async fn test_create_plan_multivector_structure() {
        use arrow_array::{Float32Array, RecordBatch};
        use arrow_schema::{DataType, Field, Schema};
        use datafusion_physical_plan::display::DisplayableExecutionPlan;
@@ -1413,18 +1037,11 @@ mod tests {
            .unwrap();
        let native_table = table.as_native().unwrap();

-        // A batch of vectors against a fixed-size vector column should use
-        // Lance's native batch KNN path instead of independent scan plans.
+        // This triggers the "create_multi_vector_plan" logic branch
        let q1 = Arc::new(Float32Array::from(vec![1.0, 2.0]));
        let q2 = Arc::new(Float32Array::from(vec![3.0, 4.0]));

        let req = VectorQueryRequest {
-            base: QueryRequest {
-                filter: Some(QueryFilter::Sql("id >= 0".to_string())),
-                limit: Some(1),
-                select: Select::Columns(vec!["id".to_string()]),
-                ..Default::default()
-            },
            column: Some("vector".to_string()),
            query_vector: vec![q1, q2],
            ..Default::default()
@@ -1441,17 +1058,19 @@ mod tests {
            .indent(true)
            .to_string();

+        // We expect a RepartitionExec wrapping a UnionExec
        assert!(
-            display.contains("KNNVectorDistance: queries=2"),
-            "plan should use native batch KNN, got:\n{display}"
+            display.contains("RepartitionExec"),
+            "Plan should include Repartitioning"
        );
        assert!(
-            !display.contains("UnionExec"),
-            "flat batch KNN should share one scan, got:\n{display}"
+            display.contains("UnionExec"),
+            "Plan should include a Union of multiple searches"
        );
+        // We expect the projection to add the 'query_index' column (logic inside multi_vector_plan)
        assert!(
            display.contains("query_index"),
-            "plan should add query_index column, got:\n{display}"
+            "Plan should add query_index column"
        );
    }

@@ -27,8 +27,6 @@ use std::sync::Arc;

 use arrow_array::Array;
 use arrow_schema::{DataType, Schema as ArrowSchema};
-use datafusion::common::{DataFusionError, ToDFSchema};
-use datafusion::prelude::SessionContext;
 use datafusion_physical_plan::expressions::Column;
 use datafusion_physical_plan::projection::ProjectionExec;
 use datafusion_physical_plan::{ExecutionPlan, PhysicalExpr};
@@ -300,7 +298,7 @@ async fn build_read_context(
    for shard_id in shard_ids {
        let manifest_store =
            ShardManifestStore::new(store.clone(), &base_path, shard_id, scan_batch_size);
-        if let Some(manifest) = manifest_store.latest().await? {
+        if let Some(manifest) = manifest_store.read_latest().await? {
            snapshots.push(snapshot_from_manifest(shard_id, &manifest, &exclude));
        }
    }
@@ -393,21 +391,7 @@ fn base_scanner(
    }
    if let Some(filter) = &query.base.filter {
        scanner = match filter {
-            QueryFilter::Sql(sql) => {
-                // Parse here instead of inside `LsmScanner::filter` so the typed
-                // DataFusion `FieldNotFound` error is still available for the
-                // same nested-field enrichment used by the ordinary scanner.
-                let schema = ArrowSchema::from(dataset.schema());
-                let df_schema = schema.clone().to_dfschema().map_err(|error| {
-                    enrich_filter_error(error, &schema, "Failed to create DFSchema")
-                })?;
-                let expr = SessionContext::new()
-                    .parse_sql_expr(sql, &df_schema)
-                    .map_err(|error| {
-                        enrich_filter_error(error, &schema, "Failed to parse filter expression")
-                    })?;
-                scanner.filter_expr(expr)
-            }
+            QueryFilter::Sql(sql) => scanner.filter(sql)?,
            QueryFilter::Datafusion(expr) => scanner.filter_expr(expr.clone()),
            QueryFilter::Substrait(_) => {
                return Err(Error::NotSupported {
@@ -419,12 +403,6 @@ fn base_scanner(
    Ok(scanner)
 }

-fn enrich_filter_error(error: DataFusionError, schema: &ArrowSchema, context: &str) -> Error {
-    super::field_not_found_diagnostic(&error, schema).unwrap_or_else(|| Error::InvalidInput {
-        message: format!("{context}: {error}"),
-    })
-}
-
 /// Plain scan: filter / projection / limit over base ∪ SSTables ∪ in-memory.
 /// The plain scan applies limit and offset inside the planner.
 async fn plain_plan(
@@ -7,16 +7,6 @@
 //! therefore idempotent and does not observe input mutation -- once a row is
 //! filled, changing what the expression reads leaves the stored result alone.
 //!
-//! A column's computed inputs are filled first -- the dependency graph is
-//! walked once, each reachable column filled once in dependency order, each
-//! fill its own commit. Every fill in the pass, the requested column's
-//! included, covers only the fragments of the snapshot the pass started
-//! from: a commit may rebase over a concurrent append, and the fragment that
-//! admits carries placeholder nulls no earlier fill covered, so it waits for
-//! a later refresh rather than being read as values. Two concurrent fills of
-//! one input collide on its field in lance's conflict check, so a dependent
-//! fill can only commit over inputs that were durable when it read them.
-//!
 //! Two passes per fragment. The first scans only the unfilled live rows and
 //! evaluates the expression over them, which yields the exact fill count and
 //! decides whether the fragment is staged at all -- a fragment where nothing
@@ -29,14 +19,10 @@
 //! inputs masked to null first, so a poison value in a row nobody is filling
 //! cannot fail the refresh.

-use std::collections::HashSet;
 use std::sync::Arc;

-use arrow_array::{
-    Array, ArrayRef, BooleanArray, LargeBinaryArray, RecordBatch, RecordBatchOptions, StructArray,
-    new_null_array,
-};
-use arrow_schema::{DataType, Field as ArrowField, Schema as ArrowSchema};
+use arrow_array::{ArrayRef, BooleanArray, RecordBatch, RecordBatchOptions};
+use arrow_schema::Schema as ArrowSchema;
 use datafusion_expr::ColumnarValue;
 use futures::{Stream, StreamExt, TryStreamExt};
 use lance::Dataset;
@@ -44,7 +30,7 @@ use lance::dataset::WriteDestination;
 use lance::dataset::fragment::FileFragment;
 use lance::dataset::transaction::Operation;
 use lance_core::ROW_ID;
-use lance_core::datatypes::{BlobHandling, Schema as LanceSchema};
+use lance_core::datatypes::Schema as LanceSchema;
 use serde::{Deserialize, Serialize};

 use super::computed_columns::{BoundExpression, ComputedColumnKind, computed_column_from_field};
@@ -55,8 +41,7 @@ use crate::{Error, Result};
 /// The result of refreshing a computed column.
 #[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
 pub struct RefreshColumnResult {
-    /// Rows that had a value computed, in the requested column only; inputs
-    /// filled on its behalf are not counted.
+    /// Rows that had a value computed.
    #[serde(default)]
    pub rows_filled: u64,
    /// The commit version associated with the operation.
@@ -67,7 +52,6 @@ pub struct RefreshColumnResult {
 struct RefreshExecution {
    result: RefreshColumnResult,
    source_version: u64,
-    published_version: Option<u64>,
 }

 /// Internal implementation of the refresh logic.
@@ -90,12 +74,7 @@ async fn execute_refresh_column_with_source(

    let expression = declared_expression(&dataset, column)?;
    let schema = Arc::new(ArrowSchema::from(dataset.schema()));
-    let bound = Arc::new(super::computed_columns::bind(
-        schema.clone(),
-        column,
-        &expression,
-    )?);
-    ensure_inputs_filled(&dataset, &schema, column, &bound).await?;
+    let bound = Arc::new(super::computed_columns::bind(schema, column, &expression)?);
    let field = dataset
        .schema()
        .field(column)
@@ -108,7 +87,6 @@ async fn execute_refresh_column_with_source(
        fields: vec![field.clone()],
        metadata: Default::default(),
    };
-    let output_is_blob = field.is_blob_v2();

    let mut rows_filled = 0u64;
    let mut replacements = Vec::new();
@@ -118,30 +96,29 @@ async fn execute_refresh_column_with_source(
            continue;
        }
        rows_filled += gained;
-        let values =
-            fill_stream(&dataset, &fragment, bound.clone(), column, output_is_blob).await?;
+        let values = fill_stream(&dataset, &fragment, bound.clone(), column).await?;
        replacements.push(fragment.write_columns(values, &column_schema).await?);
    }

-    let source_version = dataset.version().version;
    if replacements.is_empty() {
+        let source_version = dataset.version().version;
        return Ok(RefreshExecution {
            result: RefreshColumnResult {
                rows_filled: 0,
                version: source_version,
            },
            source_version,
-            published_version: None,
        });
    }

+    let read_version = dataset.version().version;
    // The dataset's own session, so registrations and caches survive the
    // commit being installed on the handle.
    let session = dataset.session();
    let new_dataset = Dataset::commit(
        WriteDestination::Dataset(dataset.clone()),
        Operation::DataReplacement { replacements },
-        Some(source_version),
+        Some(read_version),
        None,
        None,
        session,
@@ -156,52 +133,10 @@ async fn execute_refresh_column_with_source(
            rows_filled,
            version,
        },
-        source_version,
-        published_version: Some(version),
+        source_version: read_version,
    })
 }

-/// Refuse while a computed input still has rows a refresh of it would fill:
-/// read now, its placeholder null would be evaluated as a value and kept.
-async fn ensure_inputs_filled(
-    dataset: &Dataset,
-    schema: &Arc<ArrowSchema>,
-    column: &str,
-    bound: &BoundExpression,
-) -> Result<()> {
-    for input in &bound.roots {
-        let Some(declaration) = schema
-            .field_with_name(input)
-            .ok()
-            .and_then(computed_column_from_field)
-        else {
-            continue;
-        };
-        let ComputedColumnKind::Sql { expression } = &declaration.kind else {
-            return Err(Error::NotSupported {
-                message: format!(
-                    "computed column '{column}' reads '{input}', whose fill state this \
-                     refresh cannot check; refresh '{input}' first"
-                ),
-            });
-        };
-        let input_bound = super::computed_columns::bind(schema.clone(), input, expression)?;
-        let mut unfilled = 0u64;
-        for fragment in dataset.get_fragments() {
-            unfilled += count_fragment_gains(dataset, &fragment, &input_bound, input).await?;
-        }
-        if unfilled > 0 {
-            return Err(Error::InvalidInput {
-                message: format!(
-                    "computed column '{column}' reads '{input}', which has {unfilled} unfilled \
-                     rows; refresh '{input}' first"
-                ),
-            });
-        }
-    }
-    Ok(())
-}
-
 /// Run the refresh as a [`Job`] in this process.
 pub(crate) async fn execute_refresh_column_async(
    table: &NativeTable,
@@ -225,7 +160,8 @@ pub(crate) async fn execute_refresh_column_async(
            rows_failed: 0,
            rows_remaining: 0,
            source_version: execution.source_version,
-            published_version: execution.published_version,
+            published_version: (execution.result.rows_filled > 0)
+                .then_some(execution.result.version),
        })
    })))
 }
@@ -300,15 +236,12 @@ fn evaluation_batch(
    mask_out: Option<&BooleanArray>,
 ) -> lance_core::Result<RecordBatch> {
    let mut columns = Vec::with_capacity(bound.roots.len());
-    let mut fields = Vec::with_capacity(bound.roots.len());
    for name in &bound.roots {
-        let index = batch.schema_ref().index_of(name).map_err(|_| {
+        let column = batch.column_by_name(name).ok_or_else(|| {
            lance_core::Error::invalid_input(format!(
                "refreshing a computed column read no {name} column"
            ))
        })?;
-        let column = batch.column(index);
-        fields.push(batch.schema_ref().field(index).clone());
        // Rows outside the mask must not reach the expression: a value in a
        // deleted or already-filled row can be one it would choke on.
        columns.push(match mask_out {
@@ -317,7 +250,7 @@ fn evaluation_batch(
        });
    }
    Ok(RecordBatch::try_new_with_options(
-        Arc::new(ArrowSchema::new(fields)),
+        bound.read_schema.clone(),
        columns,
        &RecordBatchOptions::new().with_row_count(Some(batch.num_rows())),
    )?)
@@ -338,99 +271,6 @@ fn evaluate(bound: &BoundExpression, batch: &RecordBatch) -> lance_core::Result<
    }
 }

-fn materialized_blob_ids(schema: &LanceSchema, paths: &[String]) -> Result<HashSet<u32>> {
-    paths
-        .iter()
-        .map(|path| {
-            let field = schema
-                .resolve(path)
-                .and_then(|fields| fields.last().copied())
-                .ok_or_else(|| Error::InvalidInput {
-                    message: format!("computed Blob input '{path}' no longer exists"),
-                })?;
-            if !field.is_blob_v2() {
-                return Err(Error::InvalidInput {
-                    message: format!("computed Blob input '{path}' is no longer Blob v2"),
-                });
-            }
-            u32::try_from(field.id).map_err(|_| Error::InvalidInput {
-                message: format!(
-                    "computed Blob input '{path}' has invalid field id {}",
-                    field.id
-                ),
-            })
-        })
-        .collect()
-}
-
-fn configure_blob_inputs(
-    scanner: &mut lance::dataset::scanner::Scanner,
-    schema: &LanceSchema,
-    bound: &BoundExpression,
-    extra_blob_id: Option<u32>,
-) -> Result<()> {
-    let mut ids = materialized_blob_ids(schema, &bound.blob_paths)?;
-    ids.extend(extra_blob_id);
-    scanner.blob_handling(BlobHandling::SomeBlobsBinary(ids));
-    Ok(())
-}
-
-fn blob_array_from_binary(
-    array: &ArrayRef,
-    target_field: &ArrowField,
-) -> lance_core::Result<ArrayRef> {
-    let values = array
-        .as_any()
-        .downcast_ref::<LargeBinaryArray>()
-        .ok_or_else(|| {
-            lance_core::Error::invalid_input(format!(
-                "a Blob v2 computed output produced {}, expected LargeBinary",
-                array.data_type()
-            ))
-        })?;
-    let mut builder = lance::blob::BlobArrayBuilder::new(values.len());
-    for index in 0..values.len() {
-        if values.is_null(index) {
-            builder.push_null()?;
-        } else {
-            builder.push_bytes(values.value(index))?;
-        }
-    }
-    let minimal = builder.finish()?;
-    let minimal = minimal
-        .as_any()
-        .downcast_ref::<StructArray>()
-        .ok_or_else(|| lance_core::Error::internal("Blob builder returned a non-struct array"))?;
-    let DataType::Struct(target_fields) = target_field.data_type() else {
-        return Err(lance_core::Error::invalid_input(format!(
-            "Blob v2 output field '{}' has non-struct type {}",
-            target_field.name(),
-            target_field.data_type()
-        )));
-    };
-    let columns = target_fields
-        .iter()
-        .map(|field| match field.name().as_str() {
-            "data" | "uri" => minimal
-                .column_by_name(field.name())
-                .cloned()
-                .ok_or_else(|| {
-                    lance_core::Error::internal(format!("Blob builder omitted '{}'", field.name()))
-                }),
-            "position" | "size" => Ok(new_null_array(field.data_type(), minimal.len())),
-            name => Err(lance_core::Error::invalid_input(format!(
-                "Blob v2 output field '{}' has unsupported logical child '{name}'",
-                target_field.name()
-            ))),
-        })
-        .collect::<lance_core::Result<Vec<_>>>()?;
-    Ok(Arc::new(StructArray::try_new(
-        target_fields.clone(),
-        columns,
-        minimal.nulls().cloned(),
-    )?))
-}
-
 /// How many rows of one fragment would gain a value.
 ///
 /// Scans only the unfilled live rows -- deleted rows never reach the
@@ -449,7 +289,6 @@ async fn count_fragment_gains(
        .with_row_id()
        .filter(&format!("{} IS NULL", quote_identifier(column)))?
        .project(&bound.roots)?;
-    configure_blob_inputs(&mut scanner, dataset.schema(), bound, None)?;

    let mut gained = 0u64;
    let mut batches = scanner.try_into_stream().await?;
@@ -471,7 +310,6 @@ async fn fill_stream(
    fragment: &FileFragment,
    bound: Arc<BoundExpression>,
    column: &str,
-    output_is_blob: bool,
 ) -> Result<impl Stream<Item = lance_core::Result<RecordBatch>> + Send + use<>> {
    let mut projection: Vec<String> = bound.roots.clone();
    projection.push(column.to_string());
@@ -481,20 +319,6 @@ async fn fill_stream(
        .with_row_id()
        .include_deleted_rows()
        .project(&projection)?;
-    let output_blob_id = output_is_blob
-        .then(|| {
-            dataset
-                .schema()
-                .field(column)
-                .and_then(|field| u32::try_from(field.id).ok())
-        })
-        .flatten();
-    configure_blob_inputs(
-        &mut scanner,
-        dataset.schema(),
-        bound.as_ref(),
-        output_blob_id,
-    )?;

    let projected = Arc::new(ArrowSchema::new(vec![
        ArrowSchema::from(dataset.schema())
@@ -530,11 +354,6 @@ async fn fill_stream(

        let computed = evaluate(&bound, &evaluation_batch(&batch, &bound, Some(&keep))?)?;
        let merged = arrow_select::zip::zip(&fill, &computed, existing)?;
-        let merged = if output_is_blob {
-            blob_array_from_binary(&merged, projected.field(0))?
-        } else {
-            merged
-        };
        Ok(RecordBatch::try_new(projected.clone(), vec![merged])?)
    }))
 }
@@ -543,12 +362,8 @@ async fn fill_stream(
 mod tests {
    use std::sync::Arc;

-    use arrow_array::{
-        Array, ArrayRef, Int32Array, LargeBinaryArray, RecordBatch, StructArray, record_batch,
-    };
-    use arrow_schema::Field as ArrowField;
+    use arrow_array::{Int32Array, record_batch};
    use futures::TryStreamExt;
-    use lance_core::ROW_ID;

    use crate::connect;
    use crate::query::{ExecutableQuery, QueryBase, Select};
@@ -569,8 +384,7 @@ mod tests {
            .version)
    }

-    async fn read(table: &Table, column: &str) -> Vec<Option<i64>> {
-        use arrow_array::{Array, Int64Array};
+    async fn read(table: &Table, column: &str) -> Vec<Option<i32>> {
        let batches = table
            .query()
            .select(Select::columns(&[column]))
@@ -580,19 +394,15 @@ mod tests {
            .try_collect::<Vec<_>>()
            .await
            .unwrap();
-        let mut values: Vec<Option<i64>> = batches
+        let mut values: Vec<Option<i32>> = batches
            .iter()
            .flat_map(|batch| {
-                let array = &batch[column];
-                match array.as_any().downcast_ref::<Int32Array>() {
-                    Some(ints) => ints.iter().map(|v| v.map(i64::from)).collect::<Vec<_>>(),
-                    None => array
-                        .as_any()
-                        .downcast_ref::<Int64Array>()
-                        .unwrap()
-                        .iter()
-                        .collect::<Vec<_>>(),
-                }
+                batch[column]
+                    .as_any()
+                    .downcast_ref::<Int32Array>()
+                    .unwrap()
+                    .iter()
+                    .collect::<Vec<_>>()
            })
            .collect();
        values.sort();
@@ -604,117 +414,6 @@ mod tests {
        table.add(batch).execute().await.unwrap();
    }

-    #[test]
-    fn test_blob_output_matches_complete_logical_field() {
-        let values: ArrayRef = Arc::new(LargeBinaryArray::from(vec![
-            Some(b"hello".as_slice()),
-            None,
-        ]));
-        let field = ArrowField::new(
-            "image",
-            lance_core::datatypes::BLOB_V2_LOGICAL_TYPE.clone(),
-            true,
-        );
-
-        let output = super::blob_array_from_binary(&values, &field).unwrap();
-        assert_eq!(output.data_type(), field.data_type());
-        let output = output.as_any().downcast_ref::<StructArray>().unwrap();
-        assert_eq!(output.column_by_name("position").unwrap().null_count(), 2);
-        assert_eq!(output.column_by_name("size").unwrap().null_count(), 2);
-    }
-
-    /// The gate's reproducer: `b = coalesce(a, 0)` refreshed before `a`
-    /// must not bake zeros from `a`'s placeholder null. It is refused, and
-    /// names the input, until `a` is filled -- after every append too.
-    #[tokio::test]
-    async fn test_dependent_refresh_refuses_an_unfilled_input() {
-        let table = table_with("dependent_refresh_order", vec![1, 2, 3]).await;
-        table
-            .add_columns()
-            .computed("a", "x + 1")
-            .computed("b", "coalesce(a, 0)")
-            .execute()
-            .await
-            .unwrap();
-
-        let err = table.refresh_column("b").await.unwrap_err();
-        assert!(
-            matches!(&err, Error::InvalidInput { message } if message.contains("refresh 'a' first")),
-            "{err}"
-        );
-        assert_eq!(read(&table, "b").await, vec![None, None, None]);
-
-        assert_eq!(table.refresh_column("a").await.unwrap().rows_filled, 3);
-        assert_eq!(table.refresh_column("b").await.unwrap().rows_filled, 3);
-        assert_eq!(read(&table, "b").await, vec![Some(2), Some(3), Some(4)]);
-
-        append(&table, vec![10]).await;
-        assert!(table.refresh_column("b").await.is_err());
-        table.refresh_column("a").await.unwrap();
-        assert_eq!(table.refresh_column("b").await.unwrap().rows_filled, 1);
-        assert_eq!(
-            table.count_rows(Some("b = 0".to_string())).await.unwrap(),
-            0
-        );
-    }
-
-    /// Names that need quoting, and a nested input, survive the trip through
-    /// declaration metadata and the dependency check: the recorded inputs
-    /// are matched by name, never re-parsed as SQL.
-    #[tokio::test]
-    async fn test_dependent_refresh_handles_awkward_column_names() {
-        use arrow_array::{Int32Array, StructArray};
-        use arrow_schema::{DataType, Field, Fields};
-
-        let conn = connect("memory://").execute().await.unwrap();
-        let age_fields = Fields::from(vec![Field::new("age", DataType::Int32, true)]);
-        let meta = StructArray::new(
-            age_fields.clone(),
-            vec![Arc::new(Int32Array::from(vec![10, 20])) as _],
-            None,
-        );
-        let schema = Arc::new(arrow_schema::Schema::new(vec![
-            Field::new("camelCase", DataType::Int32, true),
-            Field::new("with-hyphen", DataType::Int32, true),
-            Field::new("meta", DataType::Struct(age_fields), true),
-        ]));
-        let batch = arrow_array::RecordBatch::try_new(
-            schema,
-            vec![
-                Arc::new(Int32Array::from(vec![1, 2])) as _,
-                Arc::new(Int32Array::from(vec![100, 200])) as _,
-                Arc::new(meta) as _,
-            ],
-        )
-        .unwrap();
-        let table = conn
-            .create_table("awkward_names", batch)
-            .execute()
-            .await
-            .unwrap();
-
-        table
-            .add_columns()
-            .computed("y", "`camelCase` * 2")
-            .computed("z", "coalesce(y, 0) + `with-hyphen` + meta.age")
-            .execute()
-            .await
-            .unwrap();
-        let z = crate::table::computed_columns::computed_columns(
-            table.schema().await.unwrap().as_ref(),
-        )
-        .into_iter()
-        .find(|c| c.name == "z")
-        .unwrap();
-        assert_eq!(z.inputs, vec!["meta.age", "with-hyphen", "y"]);
-
-        let err = table.refresh_column("z").await.unwrap_err();
-        assert!(err.to_string().contains("refresh 'y' first"), "{err}");
-        assert_eq!(table.refresh_column("y").await.unwrap().rows_filled, 2);
-        assert_eq!(table.refresh_column("z").await.unwrap().rows_filled, 2);
-        assert_eq!(read(&table, "z").await, vec![Some(112), Some(224)]);
-    }
-
    #[tokio::test]
    async fn test_refresh_fills_a_declared_column() {
        let table = table_with("refresh_fills", vec![1, 2, 3]).await;
@@ -952,8 +651,7 @@ mod tests {

        let read_back = read(&table, "doubled").await;
        assert_eq!(read_back.len(), 20_000);
-        let mut expected: Vec<Option<i64>> =
-            values.iter().map(|v| Some(i64::from(v * 2))).collect();
+        let mut expected: Vec<Option<i32>> = values.iter().map(|v| Some(v * 2)).collect();
        expected.sort();
        assert_eq!(read_back, expected);
    }
@@ -1310,366 +1008,4 @@ mod tests {
        let err = table.refresh_column("embedding").await.unwrap_err();
        assert!(matches!(err, Error::NotSupported { message } if message.contains("udf")));
    }
-
-    fn blob_batch(ids: Vec<i32>, payloads: Vec<Option<&[u8]>>) -> RecordBatch {
-        use arrow_array::Int32Array;
-        use arrow_schema::{Field, Schema};
-
-        let mut builder = lance::blob::BlobArrayBuilder::new(payloads.len());
-        for payload in payloads {
-            match payload {
-                Some(payload) => builder.push_bytes(payload).unwrap(),
-                None => builder.push_null().unwrap(),
-            }
-        }
-        RecordBatch::try_new(
-            Arc::new(Schema::new(vec![
-                Field::new("id", arrow_schema::DataType::Int32, false),
-                crate::blob("image", true),
-            ])),
-            vec![Arc::new(Int32Array::from(ids)), builder.finish().unwrap()],
-        )
-        .unwrap()
-    }
-
-    async fn create_blob_table(path: &std::path::Path, batch: RecordBatch) -> Table {
-        let conn = connect(path.to_str().unwrap()).execute().await.unwrap();
-        conn.create_table("blobs", batch).execute().await.unwrap()
-    }
-
-    #[tokio::test]
-    async fn test_refresh_inherits_and_publishes_blob_output() {
-        use arrow_array::UInt64Array;
-        use lance_arrow::{
-            BLOB_DEDICATED_SIZE_THRESHOLD_META_KEY, BLOB_INLINE_SIZE_THRESHOLD_META_KEY,
-        };
-        use lance_core::datatypes::BlobKind;
-
-        use crate::table::schema_evolution::FieldMetadataUpdate;
-
-        let tmp = tempfile::tempdir().unwrap();
-        let table = create_blob_table(
-            tmp.path(),
-            blob_batch(
-                vec![1, 2, 3, 4],
-                vec![Some(b"hello"), Some(b"ab"), Some(b""), None],
-            ),
-        )
-        .await;
-        table
-            .add_columns()
-            .computed("image_copy", "image")
-            .execute()
-            .await
-            .unwrap();
-        table
-            .update_field_metadata(&[FieldMetadataUpdate::new("image_copy")
-                .set(BLOB_INLINE_SIZE_THRESHOLD_META_KEY, "1")
-                .set(BLOB_DEDICATED_SIZE_THRESHOLD_META_KEY, "4")])
-            .await
-            .unwrap();
-
-        let first_refresh = table.refresh_column("image_copy").await.unwrap();
-        assert_eq!(first_refresh.rows_filled, 3);
-        assert_eq!(
-            table.blob_columns().await.unwrap(),
-            vec!["image".to_string(), "image_copy".to_string()]
-        );
-
-        let batches = table
-            .query()
-            .with_row_id()
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let batch = arrow_select::concat::concat_batches(&batches[0].schema(), &batches).unwrap();
-        assert!(
-            batch
-                .column_by_name("image_copy")
-                .unwrap()
-                .as_any()
-                .is::<arrow_array::StructArray>()
-        );
-        let row_ids = batch
-            .column_by_name(ROW_ID)
-            .unwrap()
-            .as_any()
-            .downcast_ref::<UInt64Array>()
-            .unwrap()
-            .values()
-            .to_vec();
-        let original = table.fetch_blobs("image", &row_ids).await.unwrap();
-        let copied = table.fetch_blobs("image_copy", &row_ids).await.unwrap();
-        assert_eq!(original, copied);
-        let ids = batch
-            .column_by_name("id")
-            .unwrap()
-            .as_any()
-            .downcast_ref::<Int32Array>()
-            .unwrap();
-        let files = table
-            .fetch_blob_files("image_copy", &row_ids)
-            .await
-            .unwrap();
-        let mut layouts = ids
-            .values()
-            .iter()
-            .copied()
-            .zip(files)
-            .map(|(id, file)| (id, file.and_then(|file| file.kind())))
-            .collect::<Vec<_>>();
-        layouts.sort_by_key(|(id, _)| *id);
-        assert_eq!(
-            layouts,
-            vec![
-                (1, Some(BlobKind::Dedicated)),
-                (2, Some(BlobKind::Packed)),
-                (3, Some(BlobKind::Inline)),
-                (4, None),
-            ]
-        );
-
-        table
-            .add(blob_batch(vec![5], vec![Some(b"appended")]))
-            .execute()
-            .await
-            .unwrap();
-        table
-            .optimize(crate::table::OptimizeAction::Compact {
-                options: crate::table::CompactionOptions::default(),
-                remap_options: None,
-            })
-            .await
-            .unwrap();
-        assert_eq!(
-            table
-                .refresh_column("image_copy")
-                .await
-                .unwrap()
-                .rows_filled,
-            1
-        );
-        assert_eq!(
-            table
-                .refresh_column("image_copy")
-                .await
-                .unwrap()
-                .rows_filled,
-            0
-        );
-
-        table.checkout(first_refresh.version).await.unwrap();
-        assert_eq!(table.count_rows(None).await.unwrap(), 4);
-        assert_eq!(
-            table.blob_columns().await.unwrap(),
-            vec!["image".to_string(), "image_copy".to_string()]
-        );
-        table.checkout_latest().await.unwrap();
-    }
-
-    #[tokio::test]
-    async fn test_refresh_inherits_nested_struct_blob_input() {
-        use arrow_array::{Int32Array, StructArray, UInt64Array};
-        use arrow_schema::{DataType, Field, Fields, Schema};
-
-        let tmp = tempfile::tempdir().unwrap();
-        let mut blob_builder = lance::blob::BlobArrayBuilder::new(2);
-        blob_builder.push_bytes(b"nested").unwrap();
-        blob_builder.push_null().unwrap();
-        let blob_field = crate::blob("image", true);
-        let metadata_fields = Fields::from(vec![blob_field.clone()]);
-        let metadata = StructArray::new(
-            metadata_fields.clone(),
-            vec![blob_builder.finish().unwrap()],
-            None,
-        );
-        let batch = RecordBatch::try_new(
-            Arc::new(Schema::new(vec![
-                Field::new("id", DataType::Int32, false),
-                Field::new("metadata", DataType::Struct(metadata_fields), true),
-            ])),
-            vec![Arc::new(Int32Array::from(vec![1, 2])), Arc::new(metadata)],
-        )
-        .unwrap();
-        let table = create_blob_table(tmp.path(), batch).await;
-        table
-            .add_columns()
-            .computed("payload_copy", "metadata.image")
-            .execute()
-            .await
-            .unwrap();
-
-        assert_eq!(
-            table
-                .refresh_column("payload_copy")
-                .await
-                .unwrap()
-                .rows_filled,
-            1
-        );
-        assert_eq!(
-            table.blob_columns().await.unwrap(),
-            vec!["metadata.image".to_string(), "payload_copy".to_string()]
-        );
-        let batches = table
-            .query()
-            .with_row_id()
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let row_ids = batches[0]
-            .column_by_name(ROW_ID)
-            .unwrap()
-            .as_any()
-            .downcast_ref::<UInt64Array>()
-            .unwrap()
-            .values();
-        let payloads = table.fetch_blobs("payload_copy", row_ids).await.unwrap();
-        assert_eq!(payloads.value(0), b"nested");
-        assert!(payloads.is_null(1));
-    }
-
-    #[tokio::test]
-    async fn test_refresh_preserves_list_shape_when_materializing_blob_input() {
-        use arrow_array::{Int32Array, ListArray};
-        use arrow_buffer::{OffsetBuffer, ScalarBuffer};
-        use arrow_schema::{DataType, Field, Schema};
-
-        let tmp = tempfile::tempdir().unwrap();
-        let mut blob_builder = lance::blob::BlobArrayBuilder::new(3);
-        blob_builder.push_bytes(b"a").unwrap();
-        blob_builder.push_bytes(b"bb").unwrap();
-        blob_builder.push_null().unwrap();
-        let item = Arc::new(crate::blob("item", true));
-        let images = ListArray::new(
-            item.clone(),
-            OffsetBuffer::new(ScalarBuffer::from(vec![0, 2, 3])),
-            blob_builder.finish().unwrap(),
-            None,
-        );
-        let batch = RecordBatch::try_new(
-            Arc::new(Schema::new(vec![
-                Field::new("id", DataType::Int32, false),
-                Field::new("images", DataType::List(item), true),
-            ])),
-            vec![Arc::new(Int32Array::from(vec![1, 2])), Arc::new(images)],
-        )
-        .unwrap();
-        let table = create_blob_table(tmp.path(), batch).await;
-        table
-            .add_columns()
-            .computed("image_payloads", "images")
-            .execute()
-            .await
-            .unwrap();
-
-        assert_eq!(
-            table
-                .refresh_column("image_payloads")
-                .await
-                .unwrap()
-                .rows_filled,
-            2
-        );
-        let batches = table
-            .query()
-            .select(Select::columns(&["image_payloads"]))
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let output = batches[0]
-            .column_by_name("image_payloads")
-            .unwrap()
-            .as_any()
-            .downcast_ref::<ListArray>()
-            .unwrap();
-        assert_eq!(output.value_offsets(), &[0, 2, 3]);
-        assert!(output.values().as_any().is::<LargeBinaryArray>());
-    }
-
-    #[tokio::test]
-    async fn test_refresh_inherits_external_blob_input() {
-        use arrow_array::{Int32Array, StringArray, UInt64Array};
-        use arrow_schema::{DataType, Field, Schema};
-
-        let tmp = tempfile::tempdir().unwrap();
-        let payload = b"external-payload";
-        let path = tmp.path().join("payload.bin");
-        std::fs::write(&path, payload).unwrap();
-        let uri = url::Url::from_file_path(path).unwrap().to_string();
-        let conn = connect(tmp.path().join("db").to_str().unwrap())
-            .execute()
-            .await
-            .unwrap();
-        let table = conn
-            .create_empty_table(
-                "external",
-                Arc::new(Schema::new(vec![
-                    Field::new("id", DataType::Int32, false),
-                    crate::blob("image", true),
-                ])),
-            )
-            .execute()
-            .await
-            .unwrap();
-        let batch = RecordBatch::try_new(
-            Arc::new(Schema::new(vec![
-                Field::new("id", DataType::Int32, false),
-                Field::new("image", DataType::Utf8, true),
-            ])),
-            vec![
-                Arc::new(Int32Array::from(vec![1])),
-                Arc::new(StringArray::from(vec![Some(uri)])),
-            ],
-        )
-        .unwrap();
-        table
-            .add(batch)
-            .allow_external_blob_outside_bases(true)
-            .execute()
-            .await
-            .unwrap();
-        table
-            .add_columns()
-            .computed("payload_copy", "image")
-            .execute()
-            .await
-            .unwrap();
-
-        assert_eq!(
-            table
-                .refresh_column("payload_copy")
-                .await
-                .unwrap()
-                .rows_filled,
-            1
-        );
-        let batches = table
-            .query()
-            .with_row_id()
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let row_ids = batches[0]
-            .column_by_name(ROW_ID)
-            .unwrap()
-            .as_any()
-            .downcast_ref::<UInt64Array>()
-            .unwrap()
-            .values();
-        let payloads = table.fetch_blobs("payload_copy", row_ids).await.unwrap();
-        assert_eq!(payloads.value(0), payload);
-    }
 }
@@ -55,9 +55,7 @@ pub struct DropColumnsResult {
 pub struct FieldMetadataUpdate {
    /// Dot-separated path to the field (e.g. `"embedding"` or `"address.zip"`).
    pub path: String,
-    /// Keys to set (`Some`) or delete (`None`). See
-    /// [`Table::update_field_metadata`](crate::Table::update_field_metadata) for
-    /// the conventional `lancedb:*` keys.
+    /// Keys to set (`Some`) or delete (`None`).
    pub metadata: HashMap<String, Option<String>>,
    /// If `true`, replace the field's entire metadata map instead of merging.
    pub replace: bool,
@@ -62,33 +62,22 @@ impl UpdateBuilder {
    }

    /// Executes the update operation.
-    pub async fn execute(mut self) -> Result<UpdateResult> {
+    pub async fn execute(self) -> Result<UpdateResult> {
        if self.columns.is_empty() {
            Err(Error::InvalidInput {
                message: "at least one column must be specified in an update operation".to_string(),
            })
        } else {
-            self.canonicalize_filter()?;
            self.parent.clone().update(self).await
        }
    }
-
-    pub(crate) fn canonicalize_filter(&mut self) -> Result<()> {
-        self.filter = self
-            .filter
-            .take()
-            .map(|predicate| crate::expr::canonicalize_sql_predicate(&predicate))
-            .transpose()?;
-        Ok(())
-    }
 }

 /// Internal implementation of the update logic
 pub(crate) async fn execute_update(
    table: &NativeTable,
-    mut update: UpdateBuilder,
+    update: UpdateBuilder,
 ) -> Result<UpdateResult> {
-    update.canonicalize_filter()?;
    table.dataset.ensure_mutable()?;

    // 1. Snapshot the current dataset
@@ -225,159 +225,6 @@ pub(crate) fn resolve_arrow_field_path(schema: &Schema, column: &str) -> Result<
    Ok((canonical_path, Field::from(*field)))
 }

-pub(crate) struct ResolvedFtsField {
-    pub canonical_path: String,
-    pub field: Field,
-    pub list_depth: usize,
-}
-
-/// Canonicalize a public FTS field path while keeping Arrow list item names hidden.
-pub(crate) fn resolve_lance_fts_field_path(
-    schema: &lance_core::datatypes::Schema,
-    column: &str,
-) -> Result<ResolvedFtsField> {
-    let names =
-        lance_core::datatypes::parse_field_path(column).map_err(|e| Error::InvalidInput {
-            message: format!("Invalid field path `{}`: {}", column, e),
-        })?;
-    let (root_name, remaining_names) = names.split_first().ok_or_else(|| Error::InvalidInput {
-        message: "FTS field path cannot be empty".to_string(),
-    })?;
-    let mut field = schema
-        .fields
-        .iter()
-        .find(|field| field.name == *root_name)
-        .or_else(|| {
-            schema
-                .fields
-                .iter()
-                .find(|field| field.name.eq_ignore_ascii_case(root_name))
-        })
-        .ok_or_else(|| fts_field_not_found(schema, column))?;
-    let mut canonical_names = vec![field.name.clone()];
-    let mut list_depth = 0;
-
-    for name in remaining_names {
-        while matches!(
-            field.data_type(),
-            DataType::List(_) | DataType::LargeList(_)
-        ) {
-            list_depth += 1;
-            field = field.children.first().ok_or_else(|| Error::Schema {
-                message: format!(
-                    "FTS field path `{}` has a list without an item field",
-                    column
-                ),
-            })?;
-        }
-        if !matches!(field.data_type(), DataType::Struct(_)) {
-            return Err(fts_field_not_found(schema, column));
-        }
-        field = field
-            .children
-            .iter()
-            .find(|field| field.name == *name)
-            .or_else(|| {
-                field
-                    .children
-                    .iter()
-                    .find(|field| field.name.eq_ignore_ascii_case(name))
-            })
-            .ok_or_else(|| fts_field_not_found(schema, column))?;
-        canonical_names.push(field.name.clone());
-    }
-
-    let mut terminal = field;
-    while matches!(
-        terminal.data_type(),
-        DataType::List(_) | DataType::LargeList(_)
-    ) {
-        list_depth += 1;
-        terminal = terminal.children.first().ok_or_else(|| Error::Schema {
-            message: format!(
-                "FTS field path `{}` has a list without an item field",
-                column
-            ),
-        })?;
-    }
-
-    let canonical_path = lance_core::datatypes::format_field_path(
-        &canonical_names
-            .iter()
-            .map(String::as_str)
-            .collect::<Vec<_>>(),
-    );
-    Ok(ResolvedFtsField {
-        canonical_path,
-        field: Field::from(field),
-        list_depth,
-    })
-}
-
-fn fts_field_not_found(schema: &lance_core::datatypes::Schema, column: &str) -> Error {
-    Error::Schema {
-        message: format!(
-            "Field path `{}` not found in schema. Available field paths: {}",
-            column,
-            schema.field_paths().join(", ")
-        ),
-    }
-}
-
-fn find_public_fts_field_path_by_id(
-    field: &lance_core::datatypes::Field,
-    field_id: i32,
-    path: &mut Vec<String>,
-) -> bool {
-    if field.id == field_id {
-        return true;
-    }
-    match field.data_type() {
-        DataType::List(_) | DataType::LargeList(_) => field
-            .children
-            .first()
-            .is_some_and(|child| find_public_fts_field_path_by_id(child, field_id, path)),
-        DataType::Struct(_) => field.children.iter().any(|child| {
-            path.push(child.name.clone());
-            let found = find_public_fts_field_path_by_id(child, field_id, path);
-            if !found {
-                path.pop();
-            }
-            found
-        }),
-        _ => false,
-    }
-}
-
-pub(crate) fn public_fts_field_path_by_id(
-    schema: &lance_core::datatypes::Schema,
-    field_id: i32,
-) -> Result<String> {
-    for root in &schema.fields {
-        let mut path = vec![root.name.clone()];
-        if find_public_fts_field_path_by_id(root, field_id, &mut path) {
-            return Ok(lance_core::datatypes::format_field_path(
-                &path.iter().map(String::as_str).collect::<Vec<_>>(),
-            ));
-        }
-    }
-    Err(Error::Schema {
-        message: format!("Field id `{}` not found in schema", field_id),
-    })
-}
-
-pub(crate) fn resolve_arrow_fts_field_path(
-    schema: &Schema,
-    column: &str,
-) -> Result<(String, Field)> {
-    let lance_schema =
-        lance_core::datatypes::Schema::try_from(schema).map_err(|e| Error::Schema {
-            message: format!("Invalid schema: {}", e),
-        })?;
-    let resolved = resolve_lance_fts_field_path(&lance_schema, column)?;
-    Ok((resolved.canonical_path, resolved.field))
-}
-
 pub fn supported_btree_data_type(dtype: &DataType) -> bool {
    dtype.is_integer()
        || dtype.is_floating()
@@ -633,36 +480,6 @@ mod tests {

    use super::*;

-    #[test]
-    fn test_public_fts_field_path_prefers_exact_case() {
-        let text_list = || {
-            DataType::List(Arc::new(Field::new(
-                "item",
-                DataType::Struct(vec![Field::new("content", DataType::Utf8, true)].into()),
-                true,
-            )))
-        };
-        let schema = Schema::new(vec![
-            Field::new("Docs", text_list(), true),
-            Field::new("docs", text_list(), true),
-        ]);
-
-        let (path, _) = resolve_arrow_fts_field_path(&schema, "docs.content").unwrap();
-        assert_eq!(path, "docs.content");
-
-        let lance_schema = lance_core::datatypes::Schema::try_from(&schema).unwrap();
-        let field_id = lance_schema
-            .resolve_case_insensitive("docs.item.content")
-            .unwrap()
-            .last()
-            .unwrap()
-            .id;
-        assert_eq!(
-            public_fts_field_path_by_id(&lance_schema, field_id).unwrap(),
-            "docs.content"
-        );
-    }
-
    #[test]
    fn test_guess_default_column() {
        let schema_no_vector = Schema::new(vec![
@@ -5,14 +5,12 @@ use std::sync::Arc;

 use arrow_array::{
    Array, ArrayRef, BinaryArray, Int64Array, LargeBinaryArray, RecordBatch, StringArray,
-    StructArray, UInt64Array, new_null_array,
+    StructArray, UInt64Array,
 };
 use arrow_schema::{DataType, Field, Fields, Schema};
 use futures::TryStreamExt;
 use lance::Dataset;
-use lance::dataset::WriteParams;
 use lance_file::version::{ConcreteFileVersion, LanceFileVersion};
-use lance_table::format::BasePath;
 use lancedb::{
    Connection, Error, Result, Table,
    blob::{BlobRangeRequest, blob},
@@ -21,7 +19,7 @@ use lancedb::{
        ListingDatabaseOptions, NewTableConfig, OPT_NEW_TABLE_ENABLE_STABLE_ROW_IDS,
    },
    query::{ExecutableQuery, QueryBase},
-    table::{AddDataMode, CompactionOptions, OptimizeAction, OptimizeStats, WriteOptions},
+    table::{AddDataMode, CompactionOptions, OptimizeAction, OptimizeStats},
 };
 use tempfile::tempdir;

@@ -263,11 +261,11 @@ async fn add_rejects_uncoercible_blob_input() -> Result<()> {
    let batch = RecordBatch::try_new(
        Arc::new(Schema::new(vec![
            Field::new("id", DataType::Int64, false),
-            Field::new("image", DataType::Int64, true),
+            Field::new("image", DataType::Utf8, true),
        ])),
        vec![
            Arc::new(Int64Array::from(vec![1])),
-            Arc::new(Int64Array::from(vec![42])),
+            Arc::new(StringArray::from(vec!["not bytes"])),
        ],
    )
    .unwrap();
@@ -1334,223 +1332,3 @@ async fn optimize_preserves_blob_v2_null_and_empty_distinction() -> Result<()> {
    );
    Ok(())
 }
-
-fn uri_struct_batch(id: i64, uri: &str) -> RecordBatch {
-    let image_field = blob("image", true);
-    let DataType::Struct(child_fields) = image_field.data_type().clone() else {
-        unreachable!("blob field is a struct");
-    };
-    let children: Vec<ArrayRef> = child_fields
-        .iter()
-        .map(|field| match field.name().as_str() {
-            "uri" => Arc::new(StringArray::from(vec![Some(uri)])) as ArrayRef,
-            _ => new_null_array(field.data_type(), 1),
-        })
-        .collect();
-    let image = StructArray::new(child_fields, children, None);
-    RecordBatch::try_new(
-        Arc::new(Schema::new(vec![
-            Field::new("id", DataType::Int64, false),
-            image_field,
-        ])),
-        vec![Arc::new(Int64Array::from(vec![id])), Arc::new(image)],
-    )
-    .unwrap()
-}
-
-fn uri_string_batch(id: i64, uri: &str) -> RecordBatch {
-    RecordBatch::try_new(
-        Arc::new(Schema::new(vec![
-            Field::new("id", DataType::Int64, false),
-            Field::new("image", DataType::Utf8, true),
-        ])),
-        vec![
-            Arc::new(Int64Array::from(vec![id])),
-            Arc::new(StringArray::from(vec![Some(uri)])),
-        ],
-    )
-    .unwrap()
-}
-
-fn write_payload_file_uri(dir: &std::path::Path, name: &str, payload: &[u8]) -> String {
-    let path = dir.join(name);
-    std::fs::write(&path, payload).unwrap();
-    url::Url::from_file_path(&path).unwrap().to_string()
-}
-
-#[tokio::test]
-async fn external_uri_struct_round_trips_with_flag() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db = connect(tmp.path().join("db").to_str().unwrap())
-        .execute()
-        .await?;
-    let payload: &[u8] = b"external-struct-payload";
-    let uri = write_payload_file_uri(tmp.path(), "payload.bin", payload);
-    let table = db
-        .create_empty_table("t", blob_table_schema())
-        .execute()
-        .await?;
-
-    table
-        .add(uri_struct_batch(1, &uri))
-        .allow_external_blob_outside_bases(true)
-        .execute()
-        .await?;
-
-    let ids = collect_row_ids(&table).await?;
-    let bytes = table.fetch_blobs("image", &ids).await?;
-    assert_eq!(bytes.value(0), payload);
-    Ok(())
-}
-
-#[tokio::test]
-async fn external_uri_add_requires_opt_in() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db = connect(tmp.path().join("db").to_str().unwrap())
-        .execute()
-        .await?;
-    let uri = write_payload_file_uri(tmp.path(), "payload.bin", b"unreachable");
-    let table = db
-        .create_empty_table("t", blob_table_schema())
-        .execute()
-        .await?;
-
-    let err = table
-        .add(uri_struct_batch(1, &uri))
-        .execute()
-        .await
-        .unwrap_err();
-
-    assert!(
-        err.to_string()
-            .contains("allow_external_blob_outside_bases"),
-        "got: {err}"
-    );
-    assert_eq!(table.count_rows(None).await?, 0);
-    Ok(())
-}
-
-#[tokio::test]
-async fn string_uri_input_round_trips_as_external_reference() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db = connect(tmp.path().join("db").to_str().unwrap())
-        .execute()
-        .await?;
-    let payload: &[u8] = b"external-string-payload";
-    let uri = write_payload_file_uri(tmp.path(), "payload.bin", payload);
-    let table = db
-        .create_empty_table("t", blob_table_schema())
-        .execute()
-        .await?;
-
-    table
-        .add(uri_string_batch(1, &uri))
-        .allow_external_blob_outside_bases(true)
-        .execute()
-        .await?;
-
-    let ids = collect_row_ids(&table).await?;
-    let bytes = table.fetch_blobs("image", &ids).await?;
-    assert_eq!(bytes.value(0), payload);
-
-    let files = table.fetch_blob_files("image", &ids).await?;
-    let file = files[0].as_ref().expect("missing blob file");
-    assert_eq!(file.uri(), Some(uri.as_str()));
-    Ok(())
-}
-
-#[tokio::test]
-async fn string_uri_inside_registered_base_does_not_need_the_flag() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db_path = tmp.path().join("db");
-    let external_base = tmp.path().join("external_base");
-    let object_dir = external_base.join("objects");
-    std::fs::create_dir_all(&object_dir).unwrap();
-    let payload: &[u8] = b"mapped-in-base";
-    let object_path = object_dir.join("mapped.bin");
-    std::fs::write(&object_path, payload).unwrap();
-    let object_uri = url::Url::from_file_path(&object_path).unwrap().to_string();
-    let base_uri = url::Url::from_file_path(&external_base)
-        .unwrap()
-        .to_string();
-
-    let db = connect(db_path.to_str().unwrap()).execute().await?;
-    let table = db
-        .create_empty_table("t", blob_table_schema())
-        .write_options(WriteOptions {
-            lance_write_params: Some(WriteParams {
-                initial_bases: Some(vec![BasePath {
-                    id: 1,
-                    name: Some("external".to_string()),
-                    path: base_uri,
-                    is_dataset_root: false,
-                }]),
-                ..Default::default()
-            }),
-        })
-        .execute()
-        .await?;
-
-    table
-        .add(uri_string_batch(1, &object_uri))
-        .execute()
-        .await?;
-
-    let ids = collect_row_ids(&table).await?;
-    let bytes = table.fetch_blobs("image", &ids).await?;
-    assert_eq!(bytes.value(0), payload);
-    Ok(())
-}
-
-#[tokio::test]
-async fn external_uri_rows_mix_with_inline_rows() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db = connect(tmp.path().join("db").to_str().unwrap())
-        .execute()
-        .await?;
-    let external_payload: &[u8] = b"external-bytes";
-    let uri = write_payload_file_uri(tmp.path(), "payload.bin", external_payload);
-    let table =
-        create_inline_blob_table(&db, "t", &[1], &[Some(b"inline-bytes".as_slice())]).await?;
-
-    table
-        .add(uri_string_batch(2, &uri))
-        .allow_external_blob_outside_bases(true)
-        .execute()
-        .await?;
-
-    let pairs = collect_id_rowid(&table).await?;
-    let row_ids: Vec<u64> = pairs.iter().map(|(_, r)| *r).collect();
-    let bytes = table.fetch_blobs("image", &row_ids).await?;
-    for (i, (id, _)) in pairs.iter().enumerate() {
-        match id {
-            1 => assert_eq!(bytes.value(i), b"inline-bytes"),
-            2 => assert_eq!(bytes.value(i), external_payload),
-            _ => unreachable!(),
-        }
-    }
-    Ok(())
-}
-
-#[tokio::test]
-async fn malformed_string_uri_is_rejected_at_write() -> Result<()> {
-    let tmp = tempdir().unwrap();
-    let db = connect(tmp.path().join("db").to_str().unwrap())
-        .execute()
-        .await?;
-    let table = db
-        .create_empty_table("t", blob_table_schema())
-        .execute()
-        .await?;
-
-    let err = table
-        .add(uri_string_batch(1, "not a uri"))
-        .allow_external_blob_outside_bases(true)
-        .execute()
-        .await
-        .unwrap_err();
-
-    assert!(err.to_string().contains("not a uri"), "got: {err}");
-    assert_eq!(table.count_rows(None).await?, 0);
-    Ok(())
-}