Bump version: 0.30.2 → 0.31.0-beta.0

## Summary
- Update Rust Lance workspace dependencies to `v5.0.0-beta.4` using
`ci/set_lance_version.py` (including lockfile refresh).
- Update Java `lance-core` dependency property to `5.0.0-beta.4` in
`java/pom.xml`.

## Verification
- `cargo clippy --workspace --tests --all-features -- -D warnings`
- `cargo fmt --all`

## Triggering tag
- https://github.com/lance-format/lance/releases/tag/v5.0.0-beta.4

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.27.0-beta.0"
+current_version = "0.27.2"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/workflows/build_linux_wheel/action.yml

@@ -23,12 +23,15 @@ runs:
   steps:
     - name: CONFIRM ARM BUILD
       shell: bash
+      env:
+        ARM_BUILD: ${{ inputs.arm-build }}
       run: |
-        echo "ARM BUILD: ${{ inputs.arm-build }}"
+        echo "ARM BUILD: $ARM_BUILD"
     - name: Build x86_64 Manylinux wheel
       if: ${{ inputs.arm-build == 'false' }}
       uses: PyO3/maturin-action@v1
       with:
+        maturin-version: "1.12.4"
         command: build
         working-directory: python
         docker-options: "-e PIP_EXTRA_INDEX_URL='https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/'"
@@ -44,6 +47,7 @@ runs:
       if: ${{ inputs.arm-build == 'true' }}
       uses: PyO3/maturin-action@v1
       with:
+        maturin-version: "1.12.4"
         command: build
         working-directory: python
         docker-options: "-e PIP_EXTRA_INDEX_URL='https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/'"

.github/workflows/build_mac_wheel/action.yml

@@ -20,6 +20,7 @@ runs:
       uses: PyO3/maturin-action@v1
       with:
         command: build
+        maturin-version: "1.12.4"
         # TODO: pass through interpreter
         args: ${{ inputs.args }}
         docker-options: "-e PIP_EXTRA_INDEX_URL='https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/'"

.github/workflows/build_windows_wheel/action.yml

@@ -25,6 +25,7 @@ runs:
       uses: PyO3/maturin-action@v1
       with:
         command: build
+        maturin-version: "1.12.4"
         args: ${{ inputs.args }}
         docker-options: "-e PIP_EXTRA_INDEX_URL='https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/'"
         working-directory: python

.github/workflows/dev.yml

@@ -15,7 +15,7 @@ jobs:
     name: Label PR
     runs-on: ubuntu-latest
     steps:
-      - uses: srvaroa/labeler@master
+      - uses: srvaroa/labeler@v1
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
   commitlint:
@@ -24,7 +24,7 @@ jobs:
     name: Verify PR title / description conforms to semantic-release
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/setup-node@v3
+      - uses: actions/setup-node@v4
         with:
           node-version: "18"
       # These rules are disabled because Github will always ensure there
@@ -47,7 +47,7 @@ jobs:
 
             ${{ github.event.pull_request.body }}
       - if: failure()
-        uses: actions/github-script@v6
+        uses: actions/github-script@v7
         with:
           script: |
             const message = `**ACTION NEEDED**

.github/workflows/docs.yml

@@ -53,7 +53,7 @@ jobs:
           python -m pip install --extra-index-url https://pypi.fury.io/lance-format/ --extra-index-url https://pypi.fury.io/lancedb/ -e .
           python -m pip install --extra-index-url https://pypi.fury.io/lance-format/ --extra-index-url https://pypi.fury.io/lancedb/ -r ../docs/requirements.txt
       - name: Set up node
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@v4
         with:
           node-version: 20
           cache: 'npm'
@@ -68,7 +68,7 @@ jobs:
         run: |
           PYTHONPATH=. mkdocs build
       - name: Setup Pages
-        uses: actions/configure-pages@v2
+        uses: actions/configure-pages@v5
       - name: Upload artifact
         uses: actions/upload-pages-artifact@v3
         with:

.github/workflows/nodejs.yml

@@ -7,6 +7,7 @@ on:
   pull_request:
     paths:
       - Cargo.toml
+      - Cargo.lock
       - nodejs/**
       - rust/**
       - docs/src/js/**
@@ -37,7 +38,7 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
-    - uses: actions/setup-node@v3
+    - uses: actions/setup-node@v4
       with:
         node-version: 20
         cache: 'npm'
@@ -77,9 +78,12 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
-    - uses: actions/setup-node@v3
+    - uses: actions/setup-node@v4
+      name: Setup Node.js 20 for build
       with:
-        node-version: ${{ matrix.node-version }}
+        # @napi-rs/cli v3 requires Node >= 20.12 (via @inquirer/prompts@8).
+        # Build always on Node 20; tests run on the matrix version below.
+        node-version: 20
         cache: 'npm'
         cache-dependency-path: nodejs/package-lock.json
     - uses: Swatinem/rust-cache@v2
@@ -87,12 +91,16 @@ jobs:
       run: |
         sudo apt update
         sudo apt install -y protobuf-compiler libssl-dev
-        npm install -g @napi-rs/cli
     - name: Build
       run: |
         npm ci --include=optional
         npm run build:debug -- --profile ci
-        npm run tsc
+    - uses: actions/setup-node@v4
+      name: Setup Node.js ${{ matrix.node-version }} for test
+      with:
+        node-version: ${{ matrix.node-version }}
+    - name: Compile TypeScript
+      run: npm run tsc
     - name: Setup localstack
       working-directory: .
       run: docker compose up --detach --wait
@@ -136,7 +144,7 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
-    - uses: actions/setup-node@v3
+    - uses: actions/setup-node@v4
       with:
         node-version: 20
         cache: 'npm'
@@ -145,7 +153,6 @@ jobs:
     - name: Install dependencies
       run: |
         brew install protobuf
-        npm install -g @napi-rs/cli
     - name: Build
       run: |
         npm ci --include=optional

.github/workflows/npm-publish.yml

@@ -19,6 +19,7 @@ on:
     paths:
       - .github/workflows/npm-publish.yml
       - Cargo.toml # Change in dependency frequently breaks builds
+      - Cargo.lock
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
@@ -124,20 +125,22 @@ jobs:
             pre_build: |-
               set -e &&
               apt-get update &&
-              apt-get install -y protobuf-compiler pkg-config
+              apt-get install -y protobuf-compiler pkg-config &&
+              # The base image (manylinux2014-cross) sets TARGET_CC to the old
+              # GCC 4.8 cross-compiler. aws-lc-sys checks TARGET_CC before CC,
+              # so it picks up GCC even though the napi-rs image sets CC=clang.
+              # Override to use the image's clang-18 which supports -fuse-ld=lld.
+              export TARGET_CC=clang TARGET_CXX=clang++
           - target: x86_64-unknown-linux-musl
             # This one seems to need some extra memory
             host: ubuntu-2404-8x-x64
-            # https://github.com/napi-rs/napi-rs/blob/main/alpine.Dockerfile
-            docker: ghcr.io/napi-rs/napi-rs/nodejs-rust:lts-alpine
             features: fp16kernels
             pre_build: |-
               set -e &&
-              apk add protobuf-dev curl &&
+              sudo apt-get update &&
-              ln -s /usr/lib/gcc/x86_64-alpine-linux-musl/14.2.0/crtbeginS.o /usr/lib/crtbeginS.o &&
+              sudo apt-get install -y protobuf-compiler pkg-config &&
-              ln -s /usr/lib/libgcc_s.so /usr/lib/libgcc.so &&
+              rustup target add x86_64-unknown-linux-musl &&
-              CC=gcc &&
+              export EXTRA_ARGS="-x"
-              CXX=g++
           - target: aarch64-unknown-linux-gnu
             host: ubuntu-2404-8x-x64
             # https://github.com/napi-rs/napi-rs/blob/main/debian-aarch64.Dockerfile
@@ -147,21 +150,20 @@ jobs:
               set -e &&
               apt-get update &&
               apt-get install -y protobuf-compiler pkg-config &&
-              # https://github.com/aws/aws-lc-rs/issues/737#issuecomment-2725918627
+              export TARGET_CC=clang TARGET_CXX=clang++ &&
-              ln -s /usr/aarch64-unknown-linux-gnu/lib/gcc/aarch64-unknown-linux-gnu/4.8.5/crtbeginS.o /usr/aarch64-unknown-linux-gnu/aarch64-unknown-linux-gnu/sysroot/usr/lib/crtbeginS.o &&
+              # The manylinux2014 sysroot has glibc 2.17 headers which lack
-              ln -s /usr/aarch64-unknown-linux-gnu/lib/gcc /usr/aarch64-unknown-linux-gnu/aarch64-unknown-linux-gnu/sysroot/usr/lib/gcc &&
+              # 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
           - target: aarch64-unknown-linux-musl
             host: ubuntu-2404-8x-x64
-            # https://github.com/napi-rs/napi-rs/blob/main/alpine.Dockerfile
-            docker: ghcr.io/napi-rs/napi-rs/nodejs-rust:lts-alpine
             features: ","
             pre_build: |-
               set -e &&
-              apk add protobuf-dev &&
+              sudo apt-get update &&
+              sudo apt-get install -y protobuf-compiler &&
               rustup target add aarch64-unknown-linux-musl &&
-              export CC_aarch64_unknown_linux_musl=aarch64-linux-musl-gcc &&
+              export EXTRA_ARGS="-x"
-              export CXX_aarch64_unknown_linux_musl=aarch64-linux-musl-g++
     name: build - ${{ matrix.settings.target }}
     runs-on: ${{ matrix.settings.host }}
     defaults:
@@ -192,12 +194,18 @@ jobs:
             .cargo-cache
             target/
           key: nodejs-${{ matrix.settings.target }}-cargo-${{ matrix.settings.host }}
-      - name: Setup toolchain
-        run: ${{ matrix.settings.setup }}
-        if: ${{ matrix.settings.setup }}
-        shell: bash
       - name: Install dependencies
         run: npm ci
+      - name: Install Zig
+        uses: mlugg/setup-zig@v2
+        if: ${{ contains(matrix.settings.target, 'musl') }}
+        with:
+          version: 0.14.1
+      - name: Install cargo-zigbuild
+        uses: taiki-e/install-action@v2
+        if: ${{ contains(matrix.settings.target, 'musl') }}
+        with:
+          tool: cargo-zigbuild
       - name: Build in docker
         uses: addnab/docker-run-action@v3
         if: ${{ matrix.settings.docker }}
@@ -210,24 +218,24 @@ jobs:
           run: |
             set -e
             ${{ matrix.settings.pre_build }}
-            npx napi build --platform  --release --no-const-enum \
+            npx napi build --platform --release \
               --features ${{ matrix.settings.features }} \
               --target ${{ matrix.settings.target }} \
               --dts ../lancedb/native.d.ts \
               --js ../lancedb/native.js \
               --strip \
-              dist/
+              --output-dir dist/
       - name: Build
         run: |
           ${{ matrix.settings.pre_build }}
-          npx napi build --platform  --release --no-const-enum \
+          npx napi build --platform --release \
               --features ${{ matrix.settings.features }} \
               --target ${{ matrix.settings.target }} \
               --dts ../lancedb/native.d.ts \
               --js ../lancedb/native.js \
               --strip \
               $EXTRA_ARGS \
-              dist/
+              --output-dir dist/
         if: ${{ !matrix.settings.docker }}
         shell: bash
       - name: Upload artifact
@@ -265,7 +273,7 @@ jobs:
           - target: x86_64-unknown-linux-gnu
             host: ubuntu-latest
           - target: aarch64-unknown-linux-gnu
-            host: buildjet-16vcpu-ubuntu-2204-arm
+            host: ubuntu-2404-8x-arm64
         node:
           - '20'
     runs-on: ${{ matrix.settings.host }}
@@ -355,7 +363,8 @@ jobs:
           if [[ $DRY_RUN == "true" ]]; then
             ARGS="$ARGS --dry-run"
           fi
-          if [[ $GITHUB_REF =~ refs/tags/v(.*)-beta.* ]]; then
+          VERSION=$(node -p "require('./package.json').version")
+          if [[ $VERSION == *-* ]]; then
             ARGS="$ARGS --tag preview"
           fi
           npm publish $ARGS

.github/workflows/pypi-publish.yml

@@ -9,6 +9,7 @@ on:
     paths:
       - .github/workflows/pypi-publish.yml
       - Cargo.toml # Change in dependency frequently breaks builds
+      - Cargo.lock
 
 env:
   PIP_EXTRA_INDEX_URL: "https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/"

.github/workflows/python.yml

@@ -7,9 +7,14 @@ on:
   pull_request:
     paths:
       - Cargo.toml
+      - Cargo.lock
       - python/**
       - rust/**
       - .github/workflows/python.yml
+      - .github/workflows/build_linux_wheel/**
+      - .github/workflows/build_mac_wheel/**
+      - .github/workflows/build_windows_wheel/**
+      - .github/workflows/run_tests/**
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}

.github/workflows/rust.yml

@@ -7,6 +7,7 @@ on:
   pull_request:
     paths:
       - Cargo.toml
+      - Cargo.lock
       - rust/**
       - .github/workflows/rust.yml
 
@@ -100,7 +101,9 @@ jobs:
           lfs: true
       - uses: Swatinem/rust-cache@v2
       - name: Install dependencies
-        run: sudo apt install -y protobuf-compiler libssl-dev
+        run: |
+          sudo apt update
+          sudo apt install -y protobuf-compiler libssl-dev
       - uses: rui314/setup-mold@v1
       - name: Make Swap
         run: |
@@ -204,14 +207,14 @@ jobs:
       - name: Downgrade  dependencies
         # These packages have newer requirements for MSRV
         run: |
-          cargo update -p aws-sdk-bedrockruntime --precise 1.64.0
+          cargo update -p aws-sdk-bedrockruntime --precise 1.77.0
-          cargo update -p aws-sdk-dynamodb --precise 1.55.0
+          cargo update -p aws-sdk-dynamodb --precise 1.68.0
-          cargo update -p aws-config --precise 1.5.10
+          cargo update -p aws-config --precise 1.6.0
-          cargo update -p aws-sdk-kms --precise 1.51.0
+          cargo update -p aws-sdk-kms --precise 1.63.0
-          cargo update -p aws-sdk-s3 --precise 1.65.0
+          cargo update -p aws-sdk-s3 --precise 1.79.0
-          cargo update -p aws-sdk-sso --precise 1.50.0
+          cargo update -p aws-sdk-sso --precise 1.62.0
-          cargo update -p aws-sdk-ssooidc --precise 1.51.0
+          cargo update -p aws-sdk-ssooidc --precise 1.63.0
-          cargo update -p aws-sdk-sts --precise 1.51.0
+          cargo update -p aws-sdk-sts --precise 1.63.0
           cargo update -p home --precise 0.5.9
       - name: cargo +${{ matrix.msrv }} check
         env:

Cargo.toml

@@ -5,7 +5,7 @@ exclude = ["python"]
 resolver = "2"
 
 [workspace.package]
-edition = "2021"
+edition = "2024"
 authors = ["LanceDB Devs <dev@lancedb.com>"]
 license = "Apache-2.0"
 repository = "https://github.com/lancedb/lancedb"
@@ -15,20 +15,20 @@ categories = ["database-implementations"]
 rust-version = "1.91.0"
 
 [workspace.dependencies]
-lance = { "version" = "=3.0.0-beta.5", default-features = false, "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance = { "version" = "=5.0.0-beta.4", default-features = false, "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-core = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-datagen = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-file = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=3.0.0-beta.5", default-features = false, "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-io = { "version" = "=5.0.0-beta.4", default-features = false, "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-index = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-linalg = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=3.0.0-beta.5", default-features = false, "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace-impls = { "version" = "=5.0.0-beta.4", default-features = false, "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-table = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-testing = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-datafusion = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-encoding = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=3.0.0-beta.5", "tag" = "v3.0.0-beta.5", "git" = "https://github.com/lance-format/lance.git" }
+lance-arrow = { "version" = "=5.0.0-beta.4", "tag" = "v5.0.0-beta.4", "git" = "https://github.com/lance-format/lance.git" }
 ahash = "0.8"
 # Note that this one does not include pyarrow
 arrow = { version = "57.2", optional = false }
@@ -40,13 +40,15 @@ arrow-schema = "57.2"
 arrow-select = "57.2"
 arrow-cast = "57.2"
 async-trait = "0"
-datafusion = { version = "51.0", default-features = false }
+datafusion = { version = "52.1", default-features = false }
-datafusion-catalog = "51.0"
+datafusion-catalog = "52.1"
-datafusion-common = { version = "51.0", default-features = false }
+datafusion-common = { version = "52.1", default-features = false }
-datafusion-execution = "51.0"
+datafusion-execution = "52.1"
-datafusion-expr = "51.0"
+datafusion-expr = "52.1"
-datafusion-physical-plan = "51.0"
+datafusion-functions = "52.1"
-datafusion-physical-expr = "51.0"
+datafusion-physical-plan = "52.1"
+datafusion-physical-expr = "52.1"
+datafusion-sql = "52.1"
 env_logger = "0.11"
 half = { "version" = "2.7.1", default-features = false, features = [
     "num-traits",

ci/check_lance_release.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import argparse
+import functools
 import json
 import os
 import re
@@ -26,6 +27,7 @@ SEMVER_RE = re.compile(
 )
 
 
+@functools.total_ordering
 @dataclass(frozen=True)
 class SemVer:
     major: int
@@ -156,7 +158,9 @@ def read_current_version(repo_root: Path) -> str:
 
 
 def determine_latest_tag(tags: Iterable[TagInfo]) -> TagInfo:
-    return max(tags, key=lambda tag: tag.semver)
+    # Stable releases (no prerelease) are always preferred over pre-releases.
+    # Within each group, standard semver ordering applies.
+    return max(tags, key=lambda tag: (not tag.semver.prerelease, tag.semver))
 
 
 def write_outputs(args: argparse.Namespace, payload: dict) -> None:

docker-compose.yml

@@ -1,7 +1,7 @@
 version: "3.9"
 services:
   localstack:
-    image: localstack/localstack:3.3
+    image: localstack/localstack:4.0
     ports:
       - 4566:4566
     environment:

dockerfiles/Dockerfile

@@ -1,27 +1,27 @@
-#Simple base dockerfile that supports basic dependencies required to run lance with FTS and Hybrid Search
+# Simple base dockerfile that supports basic dependencies required to run lance with FTS and Hybrid Search
-#Usage docker build -t lancedb:latest -f Dockerfile .
+# Usage: docker build -t lancedb:latest -f Dockerfile .
-FROM python:3.10-slim-buster
+FROM python:3.12-slim-bookworm
 
-# Install Rust
+# Install build dependencies in a single layer
-RUN apt-get update && apt-get install -y curl build-essential && \
+RUN apt-get update && \
-  curl https://sh.rustup.rs -sSf | sh -s -- -y
+  apt-get install -y --no-install-recommends \
-
+    curl \
-# Set the environment variable for Rust
+    build-essential \
-ENV PATH="/root/.cargo/bin:${PATH}"
+    protobuf-compiler \
-
+    git \
-# Install protobuf compiler
+    ca-certificates && \
-RUN apt-get install -y protobuf-compiler && \
   apt-get clean && \
   rm -rf /var/lib/apt/lists/*
 
-RUN apt-get -y update &&\
+# Install Rust (pinned installer, non-interactive)
-  apt-get -y upgrade && \
+RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain stable --profile minimal
-  apt-get -y install git
 
+# Set the environment variable for Rust
+ENV PATH="/root/.cargo/bin:${PATH}"
 
 # Verify installations
 RUN python --version && \
   rustc --version && \
   protoc --version
 
-RUN pip install tantivy lancedb
+RUN pip install --no-cache-dir tantivy lancedb

docs/mkdocs.yml

@@ -52,14 +52,21 @@ plugins:
           options:
             docstring_style: numpy
             heading_level: 3
-            show_source: true
-            show_symbol_type_in_heading: true
             show_signature_annotations: true
             show_root_heading: true
+            show_docstring_examples: true
+            show_docstring_attributes: false
+            show_docstring_other_parameters: true
+            show_symbol_type_heading: true
+            show_labels: false
+            show_if_no_docstring: true
+            show_source: false
             members_order: source
             docstring_section_style: list
             signature_crossrefs: true
             separate_signature: true
+            filters:
+              - "!^_"
           import:
             # for cross references
             - https://arrow.apache.org/docs/objects.inv
@@ -113,7 +120,7 @@ markdown_extensions:
       emoji_index: !!python/name:material.extensions.emoji.twemoji
       emoji_generator: !!python/name:material.extensions.emoji.to_svg
   - markdown.extensions.toc:
-      toc_depth: 3
+      toc_depth: 4
       permalink: true
       permalink_title: Anchor link to this section
 

docs/requirements.txt

@@ -1,9 +1,9 @@
-mkdocs==1.5.3
+mkdocs==1.6.1
 mkdocs-jupyter==0.24.1
-mkdocs-material==9.5.3
+mkdocs-material==9.6.23
-mkdocs-autorefs<=1.0
+mkdocs-autorefs>=0.5,<=1.0
-mkdocstrings[python]==0.25.2
+mkdocstrings[python]>=0.24,<1.0
-griffe
+griffe>=0.40,<1.0
-mkdocs-render-swagger-plugin
+mkdocs-render-swagger-plugin>=0.1.0
-pydantic
+pydantic>=2.0,<3.0
-mkdocs-redirects
+mkdocs-redirects>=1.2.0

docs/src/java/java.md

@@ -14,7 +14,7 @@ Add the following dependency to your `pom.xml`:
 <dependency>
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-core</artifactId>
-    <version>0.27.0-beta.0</version>
+    <version>0.27.2</version>
 </dependency>
 ```
 
@@ -57,32 +57,32 @@ LanceNamespace namespaceClient = LanceDbNamespaceClientBuilder.newBuilder()
 
 ## Metadata Operations
 
-### Creating a Namespace
+### Creating a Namespace Path
 
-Namespaces organize tables hierarchically. Create a namespace before creating tables within it:
+Namespace paths organize tables hierarchically. Create the desired namespace path before creating tables within it:
 
 ```java
 import org.lance.namespace.model.CreateNamespaceRequest;
 import org.lance.namespace.model.CreateNamespaceResponse;
 
-// Create a child namespace
+// Create a child namespace path
 CreateNamespaceRequest request = new CreateNamespaceRequest();
 request.setId(Arrays.asList("my_namespace"));
 
 CreateNamespaceResponse response = namespaceClient.createNamespace(request);
 ```
 
-You can also create nested namespaces:
+You can also create nested namespace paths:
 
 ```java
-// Create a nested namespace: parent/child
+// Create a nested namespace path: parent/child
 CreateNamespaceRequest request = new CreateNamespaceRequest();
 request.setId(Arrays.asList("parent_namespace", "child_namespace"));
 
 CreateNamespaceResponse response = namespaceClient.createNamespace(request);
 ```
 
-### Describing a Namespace
+### Describing a Namespace Path
 
 ```java
 import org.lance.namespace.model.DescribeNamespaceRequest;
@@ -95,22 +95,22 @@ DescribeNamespaceResponse response = namespaceClient.describeNamespace(request);
 System.out.println("Namespace properties: " + response.getProperties());
 ```
 
-### Listing Namespaces
+### Listing Namespace Paths
 
 ```java
 import org.lance.namespace.model.ListNamespacesRequest;
 import org.lance.namespace.model.ListNamespacesResponse;
 
-// List all namespaces at root level
+// List all namespace paths at the root level
 ListNamespacesRequest request = new ListNamespacesRequest();
 request.setId(Arrays.asList());  // Empty for root
 
 ListNamespacesResponse response = namespaceClient.listNamespaces(request);
 for (String ns : response.getNamespaces()) {
-    System.out.println("Namespace: " + ns);
+    System.out.println("Namespace path: " + ns);
 }
 
-// List child namespaces under a parent
+// List child namespace paths under a parent path
 ListNamespacesRequest childRequest = new ListNamespacesRequest();
 childRequest.setId(Arrays.asList("parent_namespace"));
 
@@ -123,7 +123,7 @@ ListNamespacesResponse childResponse = namespaceClient.listNamespaces(childReque
 import org.lance.namespace.model.ListTablesRequest;
 import org.lance.namespace.model.ListTablesResponse;
 
-// List tables in a namespace
+// List tables in a namespace path
 ListTablesRequest request = new ListTablesRequest();
 request.setId(Arrays.asList("my_namespace"));
 
@@ -133,7 +133,7 @@ for (String table : response.getTables()) {
 }
 ```
 
-### Dropping a Namespace
+### Dropping a Namespace Path
 
 ```java
 import org.lance.namespace.model.DropNamespaceRequest;
@@ -175,7 +175,7 @@ DropTableResponse response = namespaceClient.dropTable(request);
 
 ### Creating a Table
 
-Tables are created within a namespace by providing data in Apache Arrow IPC format:
+Tables are created within a namespace path by providing data in Apache Arrow IPC format:
 
 ```java
 import org.lance.namespace.LanceNamespace;
@@ -242,7 +242,7 @@ try (BufferAllocator allocator = new RootAllocator();
     }
     byte[] tableData = out.toByteArray();
 
-    // Create table in a namespace
+    // Create a table in a namespace path
     CreateTableRequest request = new CreateTableRequest();
     request.setId(Arrays.asList("my_namespace", "my_table"));
     CreateTableResponse response = namespaceClient.createTable(request, tableData);

docs/src/js/classes/Connection.md

@@ -61,8 +61,8 @@ sharing the same data, deletion, and index files.
 * **options.sourceVersion?**: `number`
     The version of the source table to clone.
 
-* **options.targetNamespace?**: `string`[]
+* **options.targetNamespacePath?**: `string`[]
-    The namespace for the target table (defaults to root namespace).
+    The namespace path for the target table (defaults to root namespace).
 
 #### Returns
 
@@ -116,13 +116,13 @@ Creates a new empty Table
 
 `Promise`<[`Table`](Table.md)>
 
-#### createEmptyTable(name, schema, namespace, options)
+#### createEmptyTable(name, schema, namespacePath, options)
 
 ```ts
 abstract createEmptyTable(
    name,
    schema,
-   namespace?,
+   namespacePath?,
    options?): Promise<Table>
 ```
 
@@ -136,8 +136,8 @@ Creates a new empty Table
 * **schema**: [`SchemaLike`](../type-aliases/SchemaLike.md)
     The schema of the table
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace to create the table in (defaults to root namespace)
+    The namespace path to create the table in (defaults to root namespace)
 
 * **options?**: `Partial`&lt;[`CreateTableOptions`](../interfaces/CreateTableOptions.md)&gt;
     Additional options
@@ -150,10 +150,10 @@ Creates a new empty Table
 
 ### createTable()
 
-#### createTable(options, namespace)
+#### createTable(options, namespacePath)
 
 ```ts
-abstract createTable(options, namespace?): Promise<Table>
+abstract createTable(options, namespacePath?): Promise<Table>
 ```
 
 Creates a new Table and initialize it with new data.
@@ -163,8 +163,8 @@ Creates a new Table and initialize it with new data.
 * **options**: `object` & `Partial`&lt;[`CreateTableOptions`](../interfaces/CreateTableOptions.md)&gt;
     The options object.
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace to create the table in (defaults to root namespace)
+    The namespace path to create the table in (defaults to root namespace)
 
 ##### Returns
 
@@ -197,13 +197,13 @@ Creates a new Table and initialize it with new data.
 
 `Promise`&lt;[`Table`](Table.md)&gt;
 
-#### createTable(name, data, namespace, options)
+#### createTable(name, data, namespacePath, options)
 
 ```ts
 abstract createTable(
    name,
    data,
-   namespace?,
+   namespacePath?,
    options?): Promise<Table>
 ```
 
@@ -218,8 +218,8 @@ Creates a new Table and initialize it with new data.
     Non-empty Array of Records
     to be inserted into the table
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace to create the table in (defaults to root namespace)
+    The namespace path to create the table in (defaults to root namespace)
 
 * **options?**: `Partial`&lt;[`CreateTableOptions`](../interfaces/CreateTableOptions.md)&gt;
     Additional options
@@ -247,15 +247,15 @@ Return a brief description of the connection
 ### dropAllTables()
 
 ```ts
-abstract dropAllTables(namespace?): Promise<void>
+abstract dropAllTables(namespacePath?): Promise<void>
 ```
 
 Drop all tables in the database.
 
 #### Parameters
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace to drop tables from (defaults to root namespace).
+    The namespace path to drop tables from (defaults to root namespace).
 
 #### Returns
 
@@ -266,7 +266,7 @@ Drop all tables in the database.
 ### dropTable()
 
 ```ts
-abstract dropTable(name, namespace?): Promise<void>
+abstract dropTable(name, namespacePath?): Promise<void>
 ```
 
 Drop an existing table.
@@ -276,8 +276,8 @@ Drop an existing table.
 * **name**: `string`
     The name of the table to drop.
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace of the table (defaults to root namespace).
+    The namespace path of the table (defaults to root namespace).
 
 #### Returns
 
@@ -304,7 +304,7 @@ Return true if the connection has not been closed
 ```ts
 abstract openTable(
    name,
-   namespace?,
+   namespacePath?,
    options?): Promise<Table>
 ```
 
@@ -315,8 +315,8 @@ Open a table in the database.
 * **name**: `string`
     The name of the table
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace of the table (defaults to root namespace)
+    The namespace path of the table (defaults to root namespace)
 
 * **options?**: `Partial`&lt;[`OpenTableOptions`](../interfaces/OpenTableOptions.md)&gt;
     Additional options
@@ -349,10 +349,10 @@ Tables will be returned in lexicographical order.
 
 `Promise`&lt;`string`[]&gt;
 
-#### tableNames(namespace, options)
+#### tableNames(namespacePath, options)
 
 ```ts
-abstract tableNames(namespace?, options?): Promise<string[]>
+abstract tableNames(namespacePath?, options?): Promise<string[]>
 ```
 
 List all the table names in this database.
@@ -361,8 +361,8 @@ Tables will be returned in lexicographical order.
 
 ##### Parameters
 
-* **namespace?**: `string`[]
+* **namespacePath?**: `string`[]
-    The namespace to list tables from (defaults to root namespace)
+    The namespace path to list tables from (defaults to root namespace)
 
 * **options?**: `Partial`&lt;[`TableNamesOptions`](../interfaces/TableNamesOptions.md)&gt;
     options to control the

docs/src/js/classes/Table.md

@@ -71,11 +71,12 @@ Add new columns with defined values.
 
 #### Parameters
 
-* **newColumnTransforms**: [`AddColumnsSql`](../interfaces/AddColumnsSql.md)[]
+* **newColumnTransforms**: `Field`<`any`> \| `Field`<`any`>[] \| `Schema`<`any`> \| [`AddColumnsSql`](../interfaces/AddColumnsSql.md)[]
-    pairs of column names and
+    Either:
-    the SQL expression to use to calculate the value of the new column. These
+    - An array of objects with column names and SQL expressions to calculate values
-    expressions will be evaluated for each row in the table, and can
+    - A single Arrow Field defining one column with its data type (column will be initialized with null values)
-    reference existing columns in the table.
+    - An array of Arrow Fields defining columns with their data types (columns will be initialized with null values)
+    - An Arrow Schema defining columns with their data types (columns will be initialized with null values)
 
 #### Returns
 
@@ -484,19 +485,7 @@ Modeled after ``VACUUM`` in PostgreSQL.
  - Prune: Removes old versions of the dataset
  - Index: Optimizes the indices, adding new data to existing indices
 
- Experimental API
+ The frequency an application should call optimize is based on the frequency of
- ----------------
-
- The optimization process is undergoing active development and may change.
- Our goal with these changes is to improve the performance of optimization and
- reduce the complexity.
-
- That being said, it is essential today to run optimize if you want the best
- performance.  It should be stable and safe to use in production, but it our
- hope that the API may be simplified (or not even need to be called) in the
- future.
-
- The frequency an application shoudl call optimize is based on the frequency of
  data modifications.  If data is frequently added, deleted, or updated then
  optimize should be run frequently.  A good rule of thumb is to run optimize if
  you have added or modified 100,000 or more records or run more than 20 data

docs/src/js/interfaces/DeleteResult.md

@@ -8,6 +8,14 @@
 
 ## Properties
 
+### numDeletedRows
+
+```ts
+numDeletedRows: number;
+```
+
+***
+
 ### version
 
 ```ts

docs/src/js/interfaces/OptimizeOptions.md

@@ -37,3 +37,12 @@ tbl.optimize({cleanupOlderThan: new Date()});
 ```ts
 deleteUnverified: boolean;
 ```
+
+Because they may be part of an in-progress transaction, files newer than
+7 days old are not deleted by default. If you are sure that there are no
+in-progress transactions, then you can set this to true to delete all
+files older than `cleanupOlderThan`.
+
+**WARNING**: This should only be set to true if you can guarantee that
+no other process is currently working on this dataset. Otherwise the
+dataset could be put into a corrupted state.

docs/src/js/namespaces/embedding/classes/EmbeddingFunction.md

@@ -52,7 +52,7 @@ new EmbeddingFunction<T, M>(): EmbeddingFunction<T, M>
 ### computeQueryEmbeddings()
 
 ```ts
-computeQueryEmbeddings(data): Promise<number[] | Float32Array | Float64Array>
+computeQueryEmbeddings(data): Promise<number[] | Uint8Array | Float32Array | Float64Array>
 ```
 
 Compute the embeddings for a single query
@@ -63,7 +63,7 @@ Compute the embeddings for a single query
 
 #### Returns
 
-`Promise`&lt;`number`[] \| `Float32Array` \| `Float64Array`&gt;
+`Promise`&lt;`number`[] \| `Uint8Array` \| `Float32Array` \| `Float64Array`&gt;
 
 ***
 

docs/src/js/namespaces/embedding/classes/TextEmbeddingFunction.md

@@ -37,7 +37,7 @@ new TextEmbeddingFunction<M>(): TextEmbeddingFunction<M>
 ### computeQueryEmbeddings()
 
 ```ts
-computeQueryEmbeddings(data): Promise<number[] | Float32Array | Float64Array>
+computeQueryEmbeddings(data): Promise<number[] | Uint8Array | Float32Array | Float64Array>
 ```
 
 Compute the embeddings for a single query
@@ -48,7 +48,7 @@ Compute the embeddings for a single query
 
 #### Returns
 
-`Promise`&lt;`number`[] \| `Float32Array` \| `Float64Array`&gt;
+`Promise`&lt;`number`[] \| `Uint8Array` \| `Float32Array` \| `Float64Array`&gt;
 
 #### Overrides
 

docs/src/js/type-aliases/IntoVector.md

@@ -7,5 +7,10 @@
 # Type Alias: IntoVector
 
 ```ts
-type IntoVector: Float32Array | Float64Array | number[] | Promise<Float32Array | Float64Array | number[]>;
+type IntoVector:
+  | Float32Array
+  | Float64Array
+  | Uint8Array
+  | number[]
+  | Promise<Float32Array | Float64Array | Uint8Array | number[]>;
 ```

docs/src/python/python.md

@@ -36,6 +36,20 @@ is also an [asynchronous API client](#connections-asynchronous).
 
 ::: lancedb.table.Tags
 
+## Expressions
+
+Type-safe expression builder for filters and projections. Use these instead
+of raw SQL strings with [where][lancedb.query.LanceQueryBuilder.where] and
+[select][lancedb.query.LanceQueryBuilder.select].
+
+::: lancedb.expr.Expr
+
+::: lancedb.expr.col
+
+::: lancedb.expr.lit
+
+::: lancedb.expr.func
+
 ## Querying (Synchronous)
 
 ::: lancedb.query.Query

java/README.md

@@ -1,4 +1,4 @@
-# LanceDB Java SDK
+# LanceDB Java Enterprise Client
 
 ## Configuration and Initialization
 

java/lancedb-core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
       <groupId>com.lancedb</groupId>
       <artifactId>lancedb-parent</artifactId>
-      <version>0.27.0-beta.0</version>
+      <version>0.27.2-final.0</version>
       <relativePath>../pom.xml</relativePath>
     </parent>
 
@@ -56,21 +56,21 @@
         <dependency>
             <groupId>org.apache.logging.log4j</groupId>
             <artifactId>log4j-slf4j2-impl</artifactId>
-            <version>2.24.3</version>
+            <version>2.25.3</version>
             <scope>test</scope>
         </dependency>
 
         <dependency>
             <groupId>org.apache.logging.log4j</groupId>
             <artifactId>log4j-core</artifactId>
-            <version>2.24.3</version>
+            <version>2.25.3</version>
             <scope>test</scope>
         </dependency>
 
         <dependency>
             <groupId>org.apache.logging.log4j</groupId>
             <artifactId>log4j-api</artifactId>
-            <version>2.24.3</version>
+            <version>2.25.3</version>
             <scope>test</scope>
         </dependency>
     </dependencies>

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.27.0-beta.0</version>
+    <version>0.27.2-final.0</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>3.0.0-beta.5</lance-core.version>
+        <lance-core.version>5.0.0-beta.4</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>
@@ -111,7 +111,7 @@
             <plugin>
                 <groupId>org.apache.maven.plugins</groupId>
                 <artifactId>maven-source-plugin</artifactId>
-                <version>2.2.1</version>
+                <version>3.3.1</version>
                 <executions>
                     <execution>
                         <id>attach-sources</id>
@@ -124,7 +124,7 @@
             <plugin>
                 <groupId>org.apache.maven.plugins</groupId>
                 <artifactId>maven-javadoc-plugin</artifactId>
-                <version>2.9.1</version>
+                <version>3.11.2</version>
                 <executions>
                     <execution>
                         <id>attach-javadocs</id>
@@ -178,15 +178,15 @@
             <plugins>
                 <plugin>
                     <artifactId>maven-clean-plugin</artifactId>
-                    <version>3.1.0</version>
+                    <version>3.4.1</version>
                 </plugin>
                 <plugin>
                     <artifactId>maven-resources-plugin</artifactId>
-                    <version>3.0.2</version>
+                    <version>3.3.1</version>
                 </plugin>
                 <plugin>
                     <artifactId>maven-compiler-plugin</artifactId>
-                    <version>3.8.1</version>
+                    <version>3.14.0</version>
                     <configuration>
                         <compilerArgs>
                             <arg>-h</arg>
@@ -205,11 +205,11 @@
                 </plugin>
                 <plugin>
                     <artifactId>maven-jar-plugin</artifactId>
-                    <version>3.0.2</version>
+                    <version>3.4.2</version>
                 </plugin>
                 <plugin>
                     <artifactId>maven-install-plugin</artifactId>
-                    <version>2.5.2</version>
+                    <version>3.1.3</version>
                 </plugin>
                 <plugin>
                     <groupId>com.diffplug.spotless</groupId>
@@ -327,7 +327,7 @@
                     <plugin>
                         <groupId>org.apache.maven.plugins</groupId>
                         <artifactId>maven-gpg-plugin</artifactId>
-                        <version>1.5</version>
+                        <version>3.2.7</version>
                         <executions>
                             <execution>
                                 <id>sign-artifacts</id>

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.27.0-beta.0"
+version = "0.27.2"
 license.workspace = true
 description.workspace = true
 repository.workspace = true
@@ -15,25 +15,27 @@ crate-type = ["cdylib"]
 async-trait.workspace = true
 arrow-ipc.workspace = true
 arrow-array.workspace = true
+arrow-buffer = "57.2"
+half.workspace = true
 arrow-schema.workspace = true
 env_logger.workspace = true
 futures.workspace = true
 lancedb = { path = "../rust/lancedb", default-features = false }
-napi = { version = "2.16.8", default-features = false, features = [
+napi = { version = "3.8.3", default-features = false, features = [
     "napi9",
     "async"
 ] }
-napi-derive = "2.16.4"
+napi-derive = "3.5.2"
 # Prevent dynamic linking of lzma, which comes from datafusion
-lzma-sys = { version = "*", features = ["static"] }
+lzma-sys = { version = "0.1", features = ["static"] }
 log.workspace = true
 
-# Workaround for build failure until we can fix it.
+# Pin to resolve build failures; update periodically for security patches.
-aws-lc-sys = "=0.28.0"
+aws-lc-sys = "=0.38.0"
-aws-lc-rs = "=1.13.0"
+aws-lc-rs = "=1.16.1"
 
 [build-dependencies]
-napi-build = "2.1"
+napi-build = "2.3.1"
 
 [features]
 default = ["remote", "lancedb/aws", "lancedb/gcs", "lancedb/azure", "lancedb/dynamodb", "lancedb/oss", "lancedb/huggingface"]

nodejs/__test__/arrow.test.ts

@@ -63,6 +63,7 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       tableFromIPC,
       DataType,
       Dictionary,
+      Uint8: ArrowUint8,
       // biome-ignore lint/suspicious/noExplicitAny: <explanation>
     } = <any>arrow;
     type Schema = ApacheArrow["Schema"];
@@ -362,6 +363,38 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         ).toEqual(new Float64().toString());
       });
 
+      it("will infer FixedSizeList<Float32> from Float32Array values", async function () {
+        const table = makeArrowTable([
+          { id: "a", vector: new Float32Array([0.1, 0.2, 0.3]) },
+          { id: "b", vector: new Float32Array([0.4, 0.5, 0.6]) },
+        ]);
+
+        expect(DataType.isFixedSizeList(table.getChild("vector")?.type)).toBe(
+          true,
+        );
+        const vectorType = table.getChild("vector")?.type;
+        expect(vectorType.listSize).toBe(3);
+        expect(vectorType.children[0].type.toString()).toEqual(
+          new Float32().toString(),
+        );
+      });
+
+      it("will infer FixedSizeList<Uint8> from Uint8Array values", async function () {
+        const table = makeArrowTable([
+          { id: "a", vector: new Uint8Array([1, 2, 3]) },
+          { id: "b", vector: new Uint8Array([4, 5, 6]) },
+        ]);
+
+        expect(DataType.isFixedSizeList(table.getChild("vector")?.type)).toBe(
+          true,
+        );
+        const vectorType = table.getChild("vector")?.type;
+        expect(vectorType.listSize).toBe(3);
+        expect(vectorType.children[0].type.toString()).toEqual(
+          new ArrowUint8().toString(),
+        );
+      });
+
       it("will use dictionary encoded strings if asked", async function () {
         const table = makeArrowTable([{ str: "hello" }]);
         expect(DataType.isUtf8(table.getChild("str")?.type)).toBe(true);

nodejs/__test__/table.test.ts

@@ -103,7 +103,7 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         },
         numIndices: 0,
         numRows: 3,
-        totalBytes: 24,
+        totalBytes: 44,
       });
     });
 
@@ -1259,6 +1259,98 @@ describe("schema evolution", function () {
     expect(await table.schema()).toEqual(expectedSchema);
   });
 
+  it("can add columns with schema for explicit data types", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Define schema for new columns with explicit data types
+    // Note: All columns must be nullable when using addColumns with Schema
+    // because they are initially populated with null values
+    const newColumnsSchema = new Schema([
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+      new Field("rating", new Int32(), true),
+    ]);
+
+    const result = await table.addColumns(newColumnsSchema);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+      new Field("rating", new Int32(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+
+    // Verify that new columns are populated with null values
+    const results = await table.query().toArray();
+    expect(results).toHaveLength(1);
+    expect(results[0].price).toBeNull();
+    expect(results[0].category).toBeNull();
+    expect(results[0].rating).toBeNull();
+  });
+
+  it("can add a single column using Field", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Add a single field
+    const priceField = new Field("price", new Float64(), true);
+    const result = await table.addColumns(priceField);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+  });
+
+  it("can add multiple columns using array of Fields", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Add multiple fields as array
+    const fields = [
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+    ];
+    const result = await table.addColumns(fields);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+  });
+
   it("can alter the columns in the schema", async function () {
     const con = await connect(tmpDir.name);
     const schema = new Schema([
@@ -1697,6 +1789,65 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(results2[0].text).toBe(data[1].text);
     });
 
+    test("full text search fast search", async () => {
+      const db = await connect(tmpDir.name);
+      const data = [{ text: "hello world", vector: [0.1, 0.2, 0.3], id: 1 }];
+      const table = await db.createTable("test", data);
+      await table.createIndex("text", {
+        config: Index.fts(),
+      });
+
+      // Insert unindexed data after creating the index.
+      await table.add([{ text: "xyz", vector: [0.4, 0.5, 0.6], id: 2 }]);
+
+      const withFlatSearch = await table
+        .search("xyz", "fts")
+        .limit(10)
+        .toArray();
+      expect(withFlatSearch.length).toBeGreaterThan(0);
+
+      const fastSearchResults = await table
+        .search("xyz", "fts")
+        .fastSearch()
+        .limit(10)
+        .toArray();
+      expect(fastSearchResults.length).toBe(0);
+
+      const nearestToTextFastSearch = await table
+        .query()
+        .nearestToText("xyz")
+        .fastSearch()
+        .limit(10)
+        .toArray();
+      expect(nearestToTextFastSearch.length).toBe(0);
+
+      // fastSearch should be chainable with other methods.
+      const chainedFastSearch = await table
+        .search("xyz", "fts")
+        .fastSearch()
+        .select(["text"])
+        .limit(5)
+        .toArray();
+      expect(chainedFastSearch.length).toBe(0);
+
+      await table.optimize();
+
+      const indexedFastSearch = await table
+        .search("xyz", "fts")
+        .fastSearch()
+        .limit(10)
+        .toArray();
+      expect(indexedFastSearch.length).toBeGreaterThan(0);
+
+      const indexedNearestToTextFastSearch = await table
+        .query()
+        .nearestToText("xyz")
+        .fastSearch()
+        .limit(10)
+        .toArray();
+      expect(indexedNearestToTextFastSearch.length).toBeGreaterThan(0);
+    });
+
     test("prewarm full text search index", async () => {
       const db = await connect(tmpDir.name);
       const data = [
@@ -2145,3 +2296,36 @@ describe("when creating an empty table", () => {
     expect((actualSchema.fields[1].type as Float64).precision).toBe(2);
   });
 });
+
+// Ensure we can create float32 arrays without using Arrow
+// by utilizing native JS TypedArray support
+//
+// https://github.com/lancedb/lancedb/issues/3115
+describe("when creating a table with Float32Array vectors", () => {
+  let tmpDir: tmp.DirResult;
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => {
+    tmpDir.removeCallback();
+  });
+
+  it("should persist Float32Array as FixedSizeList<Float32> in the LanceDB schema", async () => {
+    const db = await connect(tmpDir.name);
+    const table = await db.createTable("test", [
+      { id: "a", vector: new Float32Array([0.1, 0.2, 0.3]) },
+      { id: "b", vector: new Float32Array([0.4, 0.5, 0.6]) },
+    ]);
+
+    const schema = await table.schema();
+    const vectorField = schema.fields.find((f) => f.name === "vector");
+    expect(vectorField).toBeDefined();
+    expect(vectorField!.type).toBeInstanceOf(FixedSizeList);
+
+    const fsl = vectorField!.type as FixedSizeList;
+    expect(fsl.listSize).toBe(3);
+    expect(fsl.children[0].type.typeId).toBe(Type.Float);
+    // precision: HALF=0, SINGLE=1, DOUBLE=2
+    expect((fsl.children[0].type as Float32).precision).toBe(1);
+  });
+});

nodejs/__test__/vector_types.test.ts

@@ -0,0 +1,110 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+import * as tmp from "tmp";
+
+import { type Table, connect } from "../lancedb";
+import {
+  Field,
+  FixedSizeList,
+  Float32,
+  Int64,
+  Schema,
+  makeArrowTable,
+} from "../lancedb/arrow";
+
+describe("Vector query with different typed arrays", () => {
+  let tmpDir: tmp.DirResult;
+
+  afterEach(() => {
+    tmpDir?.removeCallback();
+  });
+
+  async function createFloat32Table(): Promise<Table> {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+    const db = await connect(tmpDir.name);
+    const schema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vec",
+        new FixedSizeList(2, new Field("item", new Float32())),
+        true,
+      ),
+    ]);
+    const data = makeArrowTable(
+      [
+        { id: 1n, vec: [1.0, 0.0] },
+        { id: 2n, vec: [0.0, 1.0] },
+        { id: 3n, vec: [1.0, 1.0] },
+      ],
+      { schema },
+    );
+    return db.createTable("test_f32", data);
+  }
+
+  it("should search with Float32Array (baseline)", async () => {
+    const table = await createFloat32Table();
+    const results = await table
+      .query()
+      .nearestTo(new Float32Array([1.0, 0.0]))
+      .limit(1)
+      .toArray();
+
+    expect(results.length).toBe(1);
+    expect(Number(results[0].id)).toBe(1);
+  });
+
+  it("should search with number[] (backward compat)", async () => {
+    const table = await createFloat32Table();
+    const results = await table
+      .query()
+      .nearestTo([1.0, 0.0])
+      .limit(1)
+      .toArray();
+
+    expect(results.length).toBe(1);
+    expect(Number(results[0].id)).toBe(1);
+  });
+
+  it("should search with Float64Array via raw path", async () => {
+    const table = await createFloat32Table();
+    const results = await table
+      .query()
+      .nearestTo(new Float64Array([1.0, 0.0]))
+      .limit(1)
+      .toArray();
+
+    expect(results.length).toBe(1);
+    expect(Number(results[0].id)).toBe(1);
+  });
+
+  it("should add multiple query vectors with Float64Array", async () => {
+    const table = await createFloat32Table();
+    const results = await table
+      .query()
+      .nearestTo(new Float64Array([1.0, 0.0]))
+      .addQueryVector(new Float64Array([0.0, 1.0]))
+      .limit(2)
+      .toArray();
+
+    expect(results.length).toBeGreaterThanOrEqual(2);
+  });
+
+  // Float16Array is only available in Node 22+; not in TypeScript's standard lib yet
+  const float16ArrayCtor = (globalThis as unknown as Record<string, unknown>)
+    .Float16Array as (new (values: number[]) => unknown) | undefined;
+  const hasFloat16 = float16ArrayCtor !== undefined;
+  const f16it = hasFloat16 ? it : it.skip;
+
+  f16it("should search with Float16Array via raw path", async () => {
+    const table = await createFloat32Table();
+    const results = await table
+      .query()
+      .nearestTo(new float16ArrayCtor!([1.0, 0.0]) as Float32Array)
+      .limit(1)
+      .toArray();
+
+    expect(results.length).toBe(1);
+    expect(Number(results[0].id)).toBe(1);
+  });
+});

nodejs/examples/package-lock.json

@@ -30,12 +30,15 @@
         "x64",
         "arm64"
       ],
+      "dev": true,
       "license": "Apache-2.0",
+      "optional": true,
       "os": [
         "darwin",
         "linux",
         "win32"
       ],
+      "peer": true,
       "dependencies": {
         "reflect-metadata": "^0.2.2"
       },
@@ -91,14 +94,15 @@
       }
     },
     "node_modules/@babel/code-frame": {
-      "version": "7.26.2",
+      "version": "7.29.0",
-      "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.26.2.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.29.0.tgz",
-      "integrity": "sha512-RJlIHRueQgwWitWgF8OdFYGZX328Ax5BCemNGlqHfplnRT9ESi8JkFlvaVYbS+UubVY6dpv87Fs2u5M29iNFVQ==",
+      "integrity": "sha512-9NhCeYjq9+3uxgdtp20LSiJXJvN0FeCtNGpJxuMFZ1Kv3cWUNb6DOhJwUvcVCzKGR66cw4njwM6hrJLqgOwbcw==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
-        "@babel/helper-validator-identifier": "^7.25.9",
+        "@babel/helper-validator-identifier": "^7.28.5",
         "js-tokens": "^4.0.0",
-        "picocolors": "^1.0.0"
+        "picocolors": "^1.1.1"
       },
       "engines": {
         "node": ">=6.9.0"
@@ -233,19 +237,21 @@
       }
     },
     "node_modules/@babel/helper-string-parser": {
-      "version": "7.25.9",
+      "version": "7.27.1",
-      "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.25.9.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.27.1.tgz",
-      "integrity": "sha512-4A/SCr/2KLd5jrtOMFzaKjVtAei3+2r/NChoBNoZ3EyP/+GlhoaEGoWOZUmFmoITP7zOJyHIMm+DYRd8o3PvHA==",
+      "integrity": "sha512-qMlSxKbpRlAridDExk92nSobyDdpPijUq2DW6oDnUqd0iOGxmQjyqhMIihI9+zv4LPyZdRje2cavWPbCbWm3eA==",
       "dev": true,
+      "license": "MIT",
       "engines": {
         "node": ">=6.9.0"
       }
     },
     "node_modules/@babel/helper-validator-identifier": {
-      "version": "7.25.9",
+      "version": "7.28.5",
-      "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.25.9.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.28.5.tgz",
-      "integrity": "sha512-Ed61U6XJc3CVRfkERJWDz4dJwKe7iLmmJsbOGu9wSloNSFttHV0I8g6UAgb7qnK5ly5bGLPd4oXZlxCdANBOWQ==",
+      "integrity": "sha512-qSs4ifwzKJSV39ucNjsvc6WVHs6b7S03sOh2OcHF9UHfVPqWWALUsNUVzhSBiItjRZoLHx7nIarVjqKVusUZ1Q==",
       "dev": true,
+      "license": "MIT",
       "engines": {
         "node": ">=6.9.0"
       }
@@ -260,25 +266,27 @@
       }
     },
     "node_modules/@babel/helpers": {
-      "version": "7.26.0",
+      "version": "7.28.6",
-      "resolved": "https://registry.npmjs.org/@babel/helpers/-/helpers-7.26.0.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/helpers/-/helpers-7.28.6.tgz",
-      "integrity": "sha512-tbhNuIxNcVb21pInl3ZSjksLCvgdZy9KwJ8brv993QtIVKJBBkYXz4q4ZbAv31GdnC+R90np23L5FbEBlthAEw==",
+      "integrity": "sha512-xOBvwq86HHdB7WUDTfKfT/Vuxh7gElQ+Sfti2Cy6yIWNW05P8iUslOVcZ4/sKbE+/jQaukQAdz/gf3724kYdqw==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
-        "@babel/template": "^7.25.9",
+        "@babel/template": "^7.28.6",
-        "@babel/types": "^7.26.0"
+        "@babel/types": "^7.28.6"
       },
       "engines": {
         "node": ">=6.9.0"
       }
     },
     "node_modules/@babel/parser": {
-      "version": "7.26.2",
+      "version": "7.29.0",
-      "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.26.2.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.29.0.tgz",
-      "integrity": "sha512-DWMCZH9WA4Maitz2q21SRKHo9QXZxkDsbNZoVD62gusNtNBBqDg9i7uOhASfTfIGNzW+O+r7+jAlM8dwphcJKQ==",
+      "integrity": "sha512-IyDgFV5GeDUVX4YdF/3CPULtVGSXXMLh1xVIgdCgxApktqnQV0r7/8Nqthg+8YLGaAtdyIlo2qIdZrbCv4+7ww==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
-        "@babel/types": "^7.26.0"
+        "@babel/types": "^7.29.0"
       },
       "bin": {
         "parser": "bin/babel-parser.js"
@@ -510,14 +518,15 @@
       }
     },
     "node_modules/@babel/template": {
-      "version": "7.25.9",
+      "version": "7.28.6",
-      "resolved": "https://registry.npmjs.org/@babel/template/-/template-7.25.9.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/template/-/template-7.28.6.tgz",
-      "integrity": "sha512-9DGttpmPvIxBb/2uwpVo3dqJ+O6RooAFOS+lB+xDqoE2PVCE8nfoHMdZLpfCQRLwvohzXISPZcgxt80xLfsuwg==",
+      "integrity": "sha512-YA6Ma2KsCdGb+WC6UpBVFJGXL58MDA6oyONbjyF/+5sBgxY/dwkhLogbMT2GXXyU84/IhRw/2D1Os1B/giz+BQ==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
-        "@babel/code-frame": "^7.25.9",
+        "@babel/code-frame": "^7.28.6",
-        "@babel/parser": "^7.25.9",
+        "@babel/parser": "^7.28.6",
-        "@babel/types": "^7.25.9"
+        "@babel/types": "^7.28.6"
       },
       "engines": {
         "node": ">=6.9.0"
@@ -542,13 +551,14 @@
       }
     },
     "node_modules/@babel/types": {
-      "version": "7.26.0",
+      "version": "7.29.0",
-      "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.26.0.tgz",
+      "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.29.0.tgz",
-      "integrity": "sha512-Z/yiTPj+lDVnF7lWeKCIJzaIkI0vYO87dMpZ4bg4TDrFe4XXLFWL1TbXU27gBP3QccxV9mZICCrnjnYlJjXHOA==",
+      "integrity": "sha512-LwdZHpScM4Qz8Xw2iKSzS+cfglZzJGvofQICy7W7v4caru4EaAmyUuO6BGrbyQ2mYV11W0U8j5mBhd14dd3B0A==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
-        "@babel/helper-string-parser": "^7.25.9",
+        "@babel/helper-string-parser": "^7.27.1",
-        "@babel/helper-validator-identifier": "^7.25.9"
+        "@babel/helper-validator-identifier": "^7.28.5"
       },
       "engines": {
         "node": ">=6.9.0"
@@ -1151,95 +1161,6 @@
         "url": "https://opencollective.com/libvips"
       }
     },
-    "node_modules/@isaacs/cliui": {
-      "version": "8.0.2",
-      "resolved": "https://registry.npmjs.org/@isaacs/cliui/-/cliui-8.0.2.tgz",
-      "integrity": "sha512-O8jcjabXaleOG9DQ0+ARXWZBTfnP4WNAqzuiJK7ll44AmxGKv/J2M4TPjxjY3znBCfvBXFzucm1twdyFybFqEA==",
-      "dependencies": {
-        "string-width": "^5.1.2",
-        "string-width-cjs": "npm:string-width@^4.2.0",
-        "strip-ansi": "^7.0.1",
-        "strip-ansi-cjs": "npm:strip-ansi@^6.0.1",
-        "wrap-ansi": "^8.1.0",
-        "wrap-ansi-cjs": "npm:wrap-ansi@^7.0.0"
-      },
-      "engines": {
-        "node": ">=12"
-      }
-    },
-    "node_modules/@isaacs/cliui/node_modules/ansi-regex": {
-      "version": "6.1.0",
-      "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-6.1.0.tgz",
-      "integrity": "sha512-7HSX4QQb4CspciLpVFwyRe79O3xsIZDDLER21kERQ71oaPodF8jL725AgJMFAYbooIqolJoRLuM81SpeUkpkvA==",
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/chalk/ansi-regex?sponsor=1"
-      }
-    },
-    "node_modules/@isaacs/cliui/node_modules/ansi-styles": {
-      "version": "6.2.1",
-      "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-6.2.1.tgz",
-      "integrity": "sha512-bN798gFfQX+viw3R7yrGWRqnrN2oRkEkUjjl4JNn4E8GxxbjtG3FbrEIIY3l8/hrwUwIeCZvi4QuOTP4MErVug==",
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/chalk/ansi-styles?sponsor=1"
-      }
-    },
-    "node_modules/@isaacs/cliui/node_modules/emoji-regex": {
-      "version": "9.2.2",
-      "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-9.2.2.tgz",
-      "integrity": "sha512-L18DaJsXSUk2+42pv8mLs5jJT2hqFkFE4j21wOmgbUqsZ2hL72NsUU785g9RXgo3s0ZNgVl42TiHp3ZtOv/Vyg=="
-    },
-    "node_modules/@isaacs/cliui/node_modules/string-width": {
-      "version": "5.1.2",
-      "resolved": "https://registry.npmjs.org/string-width/-/string-width-5.1.2.tgz",
-      "integrity": "sha512-HnLOCR3vjcY8beoNLtcjZ5/nxn2afmME6lhrDrebokqMap+XbeW8n9TXpPDOqdGK5qcI3oT0GKTW6wC7EMiVqA==",
-      "dependencies": {
-        "eastasianwidth": "^0.2.0",
-        "emoji-regex": "^9.2.2",
-        "strip-ansi": "^7.0.1"
-      },
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/sindresorhus"
-      }
-    },
-    "node_modules/@isaacs/cliui/node_modules/strip-ansi": {
-      "version": "7.1.0",
-      "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-7.1.0.tgz",
-      "integrity": "sha512-iq6eVVI64nQQTRYq2KtEg2d2uU7LElhTJwsH4YzIHZshxlgZms/wIc4VoDQTlG/IvVIrBKG06CrZnp0qv7hkcQ==",
-      "dependencies": {
-        "ansi-regex": "^6.0.1"
-      },
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/chalk/strip-ansi?sponsor=1"
-      }
-    },
-    "node_modules/@isaacs/cliui/node_modules/wrap-ansi": {
-      "version": "8.1.0",
-      "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-8.1.0.tgz",
-      "integrity": "sha512-si7QWI6zUMq56bESFvagtmzMdGOtoxfR+Sez11Mobfc7tm+VkUckk9bW2UeffTGVUbOksxmSw0AA2gs8g71NCQ==",
-      "dependencies": {
-        "ansi-styles": "^6.1.0",
-        "string-width": "^5.0.1",
-        "strip-ansi": "^7.0.1"
-      },
-      "engines": {
-        "node": ">=12"
-      },
-      "funding": {
-        "url": "https://github.com/chalk/wrap-ansi?sponsor=1"
-      }
-    },
     "node_modules/@isaacs/fs-minipass": {
       "version": "4.0.1",
       "resolved": "https://registry.npmjs.org/@isaacs/fs-minipass/-/fs-minipass-4.0.1.tgz",
@@ -1606,15 +1527,6 @@
       "resolved": "../dist",
       "link": true
     },
-    "node_modules/@pkgjs/parseargs": {
-      "version": "0.11.0",
-      "resolved": "https://registry.npmjs.org/@pkgjs/parseargs/-/parseargs-0.11.0.tgz",
-      "integrity": "sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg==",
-      "optional": true,
-      "engines": {
-        "node": ">=14"
-      }
-    },
     "node_modules/@protobufjs/aspromise": {
       "version": "1.1.2",
       "resolved": "https://registry.npmjs.org/@protobufjs/aspromise/-/aspromise-1.1.2.tgz",
@@ -1846,6 +1758,7 @@
       "version": "5.0.1",
       "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz",
       "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==",
+      "dev": true,
       "engines": {
         "node": ">=8"
       }
@@ -1854,6 +1767,7 @@
       "version": "4.3.0",
       "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz",
       "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==",
+      "dev": true,
       "dependencies": {
         "color-convert": "^2.0.1"
       },
@@ -2019,13 +1933,15 @@
     "node_modules/balanced-match": {
       "version": "1.0.2",
       "resolved": "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.2.tgz",
-      "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw=="
+      "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==",
+      "dev": true
     },
     "node_modules/brace-expansion": {
-      "version": "1.1.11",
+      "version": "1.1.12",
-      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.11.tgz",
+      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.12.tgz",
-      "integrity": "sha512-iCuPHDFgrHX7H2vEI/5xpz07zSHB00TpugqhmYtVmMO6518mCuRMoOYFldEBl0g187ufozdaHgWKcYFb61qGiA==",
+      "integrity": "sha512-9T9UjW3r0UW5c1Q7GTwllptXwhvYmEzFhzMfZ9H7FQWt+uZePjZPjBP/W1ZEyZ1twGWom5/56TF4lPcqjnDHcg==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
         "balanced-match": "^1.0.0",
         "concat-map": "0.0.1"
@@ -2102,6 +2018,19 @@
       "integrity": "sha512-E+XQCRwSbaaiChtv6k6Dwgc+bx+Bs6vuKJHHl5kox/BaKbhiXzqQOwK4cO22yElGp2OCmjwVhT3HmxgyPGnJfQ==",
       "dev": true
     },
+    "node_modules/call-bind-apply-helpers": {
+      "version": "1.0.2",
+      "resolved": "https://registry.npmjs.org/call-bind-apply-helpers/-/call-bind-apply-helpers-1.0.2.tgz",
+      "integrity": "sha512-Sp1ablJ0ivDkSzjcaJdxEunN5/XvksFJ2sMBFfq6x0ryhQV/2b/KwFe21cMpmHtPOSij8K99/wSfoEuTObmuMQ==",
+      "license": "MIT",
+      "dependencies": {
+        "es-errors": "^1.3.0",
+        "function-bind": "^1.1.2"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
     "node_modules/callsites": {
       "version": "3.1.0",
       "resolved": "https://registry.npmjs.org/callsites/-/callsites-3.1.0.tgz",
@@ -2298,9 +2227,11 @@
       }
     },
     "node_modules/cross-spawn": {
-      "version": "7.0.3",
+      "version": "7.0.6",
-      "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.3.tgz",
+      "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz",
-      "integrity": "sha512-iRDPJKUPVEND7dHPO8rkbOnPpyDygcDFtWjpeWNCgy8WP2rXcxXL8TskReQl6OrB2G7+UJrags1q15Fudc7G6w==",
+      "integrity": "sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==",
+      "dev": true,
+      "license": "MIT",
       "dependencies": {
         "path-key": "^3.1.0",
         "shebang-command": "^2.0.0",
@@ -2384,10 +2315,19 @@
         "node": "^14.15.0 || ^16.10.0 || >=18.0.0"
       }
     },
-    "node_modules/eastasianwidth": {
+    "node_modules/dunder-proto": {
-      "version": "0.2.0",
+      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/eastasianwidth/-/eastasianwidth-0.2.0.tgz",
+      "resolved": "https://registry.npmjs.org/dunder-proto/-/dunder-proto-1.0.1.tgz",
-      "integrity": "sha512-I88TYZWc9XiYHRQ4/3c5rjjfgkjhLyW2luGIheGERbNQ6OY7yTybanSpDXZa8y7VUP9YmDcYa+eyq4ca7iLqWA=="
+      "integrity": "sha512-KIN/nDJBQRcXw0MLVhZE9iQHmG68qAVIBg9CqmUYjmQIhgij9U5MFvrqkUL5FbtyyzZuOeOt0zdeRe4UY7ct+A==",
+      "license": "MIT",
+      "dependencies": {
+        "call-bind-apply-helpers": "^1.0.1",
+        "es-errors": "^1.3.0",
+        "gopd": "^1.2.0"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      }
     },
     "node_modules/ejs": {
       "version": "3.1.10",
@@ -2425,7 +2365,8 @@
     "node_modules/emoji-regex": {
       "version": "8.0.0",
       "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz",
-      "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A=="
+      "integrity": "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==",
+      "dev": true
     },
     "node_modules/error-ex": {
       "version": "1.3.2",
@@ -2442,6 +2383,51 @@
       "integrity": "sha512-zz06S8t0ozoDXMG+ube26zeCTNXcKIPJZJi8hBrF4idCLms4CG9QtK7qBl1boi5ODzFpjswb5JPmHCbMpjaYzg==",
       "dev": true
     },
+    "node_modules/es-define-property": {
+      "version": "1.0.1",
+      "resolved": "https://registry.npmjs.org/es-define-property/-/es-define-property-1.0.1.tgz",
+      "integrity": "sha512-e3nRfgfUZ4rNGL232gUgX06QNyyez04KdjFrF+LTRoOXmrOgFKDg4BCdsjW8EnT69eqdYGmRpJwiPVYNrCaW3g==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
+    "node_modules/es-errors": {
+      "version": "1.3.0",
+      "resolved": "https://registry.npmjs.org/es-errors/-/es-errors-1.3.0.tgz",
+      "integrity": "sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
+    "node_modules/es-object-atoms": {
+      "version": "1.1.1",
+      "resolved": "https://registry.npmjs.org/es-object-atoms/-/es-object-atoms-1.1.1.tgz",
+      "integrity": "sha512-FGgH2h8zKNim9ljj7dankFPcICIK9Cp5bm+c2gQSYePhpaG5+esrLODihIorn+Pe6FGJzWhXQotPv73jTaldXA==",
+      "license": "MIT",
+      "dependencies": {
+        "es-errors": "^1.3.0"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
+    "node_modules/es-set-tostringtag": {
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/es-set-tostringtag/-/es-set-tostringtag-2.1.0.tgz",
+      "integrity": "sha512-j6vWzfrGVfyXxge+O0x5sh6cvxAog0a/4Rdd2K36zCMV5eJ+/+tOAngRO8cODMNWbVRdVlmGZQL2YS3yR8bIUA==",
+      "license": "MIT",
+      "dependencies": {
+        "es-errors": "^1.3.0",
+        "get-intrinsic": "^1.2.6",
+        "has-tostringtag": "^1.0.2",
+        "hasown": "^2.0.2"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
     "node_modules/escalade": {
       "version": "3.2.0",
       "resolved": "https://registry.npmjs.org/escalade/-/escalade-3.2.0.tgz",
@@ -2554,19 +2540,21 @@
       }
     },
     "node_modules/filelist/node_modules/brace-expansion": {
-      "version": "2.0.1",
+      "version": "2.0.2",
-      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz",
+      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.2.tgz",
-      "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==",
+      "integrity": "sha512-Jt0vHyM+jmUBqojB7E1NIYadt0vI0Qxjxd2TErW94wDz+E2LAm5vKMXXwg6ZZBTHPuUlDgQHKXvjGBdfcF1ZDQ==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
         "balanced-match": "^1.0.0"
       }
     },
     "node_modules/filelist/node_modules/minimatch": {
-      "version": "5.1.6",
+      "version": "5.1.9",
-      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-5.1.6.tgz",
+      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-5.1.9.tgz",
-      "integrity": "sha512-lKwV/1brpG6mBUFHtb7NUmtABCb2WZZmm2wNiOA5hAb8VdCS4B3dtMWyvcoViccwAW/COERjXLt0zP1zXUN26g==",
+      "integrity": "sha512-7o1wEA2RyMP7Iu7GNba9vc0RWWGACJOCZBJX2GJWip0ikV+wcOsgVuY9uE8CPiyQhkGFSlhuSkZPavN7u1c2Fw==",
       "dev": true,
+      "license": "ISC",
       "dependencies": {
         "brace-expansion": "^2.0.1"
       },
@@ -2604,39 +2592,16 @@
       "resolved": "https://registry.npmjs.org/flatbuffers/-/flatbuffers-1.12.0.tgz",
       "integrity": "sha512-c7CZADjRcl6j0PlvFy0ZqXQ67qSEZfrVPynmnL+2zPc+NtMvrF8Y0QceMo7QqnSPc7+uWjUIAbvCQ5WIKlMVdQ=="
     },
-    "node_modules/foreground-child": {
-      "version": "3.3.0",
-      "resolved": "https://registry.npmjs.org/foreground-child/-/foreground-child-3.3.0.tgz",
-      "integrity": "sha512-Ld2g8rrAyMYFXBhEqMz8ZAHBi4J4uS1i/CxGMDnjyFWddMXLVcDp051DZfu+t7+ab7Wv6SMqpWmyFIj5UbfFvg==",
-      "dependencies": {
-        "cross-spawn": "^7.0.0",
-        "signal-exit": "^4.0.1"
-      },
-      "engines": {
-        "node": ">=14"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/foreground-child/node_modules/signal-exit": {
-      "version": "4.1.0",
-      "resolved": "https://registry.npmjs.org/signal-exit/-/signal-exit-4.1.0.tgz",
-      "integrity": "sha512-bzyZ1e88w9O1iNJbKnOlvYTrWPDl46O1bG0D3XInv+9tkPrxrN8jUUTiFlDkkmKWgn1M6CfIA13SuGqOa9Korw==",
-      "engines": {
-        "node": ">=14"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
     "node_modules/form-data": {
-      "version": "4.0.1",
+      "version": "4.0.5",
-      "resolved": "https://registry.npmjs.org/form-data/-/form-data-4.0.1.tgz",
+      "resolved": "https://registry.npmjs.org/form-data/-/form-data-4.0.5.tgz",
-      "integrity": "sha512-tzN8e4TX8+kkxGPK8D5u0FNmjPUjw3lwC9lSLxxoB/+GtsJG91CO8bSWy73APlgAZzZbXEYZJuxjkHH2w+Ezhw==",
+      "integrity": "sha512-8RipRLol37bNs2bhoV67fiTEvdTrbMUYcFTiy3+wuuOnUog2QBHCZWXDRijWQfAkhBj2Uf5UnVaiWwA5vdd82w==",
+      "license": "MIT",
       "dependencies": {
         "asynckit": "^0.4.0",
         "combined-stream": "^1.0.8",
+        "es-set-tostringtag": "^2.1.0",
+        "hasown": "^2.0.2",
         "mime-types": "^2.1.12"
       },
       "engines": {
@@ -2684,7 +2649,6 @@
       "version": "1.1.2",
       "resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.2.tgz",
       "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==",
-      "dev": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -2707,6 +2671,30 @@
         "node": "6.* || 8.* || >= 10.*"
       }
     },
+    "node_modules/get-intrinsic": {
+      "version": "1.3.0",
+      "resolved": "https://registry.npmjs.org/get-intrinsic/-/get-intrinsic-1.3.0.tgz",
+      "integrity": "sha512-9fSjSaos/fRIVIp+xSJlE6lfwhES7LNtKaCBIamHsjr2na1BiABJPo0mOjjz8GJDURarmCPGqaiVg5mfjb98CQ==",
+      "license": "MIT",
+      "dependencies": {
+        "call-bind-apply-helpers": "^1.0.2",
+        "es-define-property": "^1.0.1",
+        "es-errors": "^1.3.0",
+        "es-object-atoms": "^1.1.1",
+        "function-bind": "^1.1.2",
+        "get-proto": "^1.0.1",
+        "gopd": "^1.2.0",
+        "has-symbols": "^1.1.0",
+        "hasown": "^2.0.2",
+        "math-intrinsics": "^1.1.0"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/ljharb"
+      }
+    },
     "node_modules/get-package-type": {
       "version": "0.1.0",
       "resolved": "https://registry.npmjs.org/get-package-type/-/get-package-type-0.1.0.tgz",
@@ -2716,6 +2704,19 @@
         "node": ">=8.0.0"
       }
     },
+    "node_modules/get-proto": {
+      "version": "1.0.1",
+      "resolved": "https://registry.npmjs.org/get-proto/-/get-proto-1.0.1.tgz",
+      "integrity": "sha512-sTSfBjoXBp89JvIKIefqw7U2CCebsc74kiY6awiGogKtoSGbgjYE/G/+l9sF3MWFPNc9IcoOC4ODfKHfxFmp0g==",
+      "license": "MIT",
+      "dependencies": {
+        "dunder-proto": "^1.0.1",
+        "es-object-atoms": "^1.0.0"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
     "node_modules/get-stream": {
       "version": "6.0.1",
       "resolved": "https://registry.npmjs.org/get-stream/-/get-stream-6.0.1.tgz",
@@ -2758,6 +2759,18 @@
         "node": ">=4"
       }
     },
+    "node_modules/gopd": {
+      "version": "1.2.0",
+      "resolved": "https://registry.npmjs.org/gopd/-/gopd-1.2.0.tgz",
+      "integrity": "sha512-ZUKRh6/kUFoAiTAtTYPZJ3hw9wNxx+BIBOijnlG9PnrJsCcSjs1wyyD6vJpaYtgnzDrKYRSqf3OO6Rfa93xsRg==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/ljharb"
+      }
+    },
     "node_modules/graceful-fs": {
       "version": "4.2.11",
       "resolved": "https://registry.npmjs.org/graceful-fs/-/graceful-fs-4.2.11.tgz",
@@ -2778,11 +2791,37 @@
         "node": ">=8"
       }
     },
+    "node_modules/has-symbols": {
+      "version": "1.1.0",
+      "resolved": "https://registry.npmjs.org/has-symbols/-/has-symbols-1.1.0.tgz",
+      "integrity": "sha512-1cDNdwJ2Jaohmb3sg4OmKaMBwuC48sYni5HUw2DvsC8LjGTLK9h+eb1X6RyuOHe4hT0ULCW68iomhjUoKUqlPQ==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/ljharb"
+      }
+    },
+    "node_modules/has-tostringtag": {
+      "version": "1.0.2",
+      "resolved": "https://registry.npmjs.org/has-tostringtag/-/has-tostringtag-1.0.2.tgz",
+      "integrity": "sha512-NqADB8VjPFLM2V0VvHUewwwsw0ZWBaIdgo+ieHtK3hasLz4qeCRjYcqfB6AQrBggRKppKF8L52/VqdVsO47Dlw==",
+      "license": "MIT",
+      "dependencies": {
+        "has-symbols": "^1.0.3"
+      },
+      "engines": {
+        "node": ">= 0.4"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/ljharb"
+      }
+    },
     "node_modules/hasown": {
       "version": "2.0.2",
       "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.2.tgz",
       "integrity": "sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==",
-      "dev": true,
       "dependencies": {
         "function-bind": "^1.1.2"
       },
@@ -2882,6 +2921,7 @@
       "version": "3.0.0",
       "resolved": "https://registry.npmjs.org/is-fullwidth-code-point/-/is-fullwidth-code-point-3.0.0.tgz",
       "integrity": "sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==",
+      "dev": true,
       "engines": {
         "node": ">=8"
       }
@@ -2919,7 +2959,8 @@
     "node_modules/isexe": {
       "version": "2.0.0",
       "resolved": "https://registry.npmjs.org/isexe/-/isexe-2.0.0.tgz",
-      "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw=="
+      "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==",
+      "dev": true
     },
     "node_modules/istanbul-lib-coverage": {
       "version": "3.2.2",
@@ -2987,20 +3028,6 @@
         "node": ">=8"
       }
     },
-    "node_modules/jackspeak": {
-      "version": "3.4.3",
-      "resolved": "https://registry.npmjs.org/jackspeak/-/jackspeak-3.4.3.tgz",
-      "integrity": "sha512-OGlZQpz2yfahA/Rd1Y8Cd9SIEsqvXkLVoSw/cgwhnhFMDbsQFeZYoJJ7bIZBS9BcamUW96asq/npPWugM+RQBw==",
-      "dependencies": {
-        "@isaacs/cliui": "^8.0.2"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      },
-      "optionalDependencies": {
-        "@pkgjs/parseargs": "^0.11.0"
-      }
-    },
     "node_modules/jake": {
       "version": "10.9.2",
       "resolved": "https://registry.npmjs.org/jake/-/jake-10.9.2.tgz",
@@ -3605,10 +3632,11 @@
       "dev": true
     },
     "node_modules/js-yaml": {
-      "version": "3.14.1",
+      "version": "3.14.2",
-      "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.14.1.tgz",
+      "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-3.14.2.tgz",
-      "integrity": "sha512-okMH7OXXJ7YrN9Ok3/SXrnu4iX9yOk+25nqX4imS2npuvTYDmo/QEZoqwZkYaIDk3jVvBOTOIEgEhaLOynBS9g==",
+      "integrity": "sha512-PMSmkqxr106Xa156c2M265Z+FTrPl+oxd/rgOQy2tijQeK5TxQ43psO1ZCwhVOSdnn+RzkzlRz/eY4BgJBYVpg==",
       "dev": true,
+      "license": "MIT",
       "dependencies": {
         "argparse": "^1.0.7",
         "esprima": "^4.0.0"
@@ -3728,6 +3756,15 @@
         "tmpl": "1.0.5"
       }
     },
+    "node_modules/math-intrinsics": {
+      "version": "1.1.0",
+      "resolved": "https://registry.npmjs.org/math-intrinsics/-/math-intrinsics-1.1.0.tgz",
+      "integrity": "sha512-/IXtbwEk5HTPyEwyKX6hGkYXxM9nbj64B+ilVJnC/R6B0pH5G4V3b0pVbL7DBj4tkhBAppbQUlf6F6Xl9LHu1g==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 0.4"
+      }
+    },
     "node_modules/merge-stream": {
       "version": "2.0.0",
       "resolved": "https://registry.npmjs.org/merge-stream/-/merge-stream-2.0.0.tgz",
@@ -3776,10 +3813,11 @@
       }
     },
     "node_modules/minimatch": {
-      "version": "3.1.2",
+      "version": "3.1.5",
-      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-3.1.2.tgz",
+      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-3.1.5.tgz",
-      "integrity": "sha512-J7p63hRiAjw1NDEww1W7i37+ByIrOWO5XQQAzZ3VOcL0PNybwpfmV/N05zFAzwQ9USyEcX6t3UO+K5aqBQOIHw==",
+      "integrity": "sha512-VgjWUsnnT6n+NUk6eZq77zeFdpW2LWDzP6zFGrCbHXiYNul5Dzqk2HHQ5uFH2DNW5Xbp8+jVzaeNt94ssEEl4w==",
       "dev": true,
+      "license": "ISC",
       "dependencies": {
         "brace-expansion": "^1.1.7"
       },
@@ -3796,31 +3834,17 @@
       }
     },
     "node_modules/minizlib": {
-      "version": "3.0.1",
+      "version": "3.1.0",
-      "resolved": "https://registry.npmjs.org/minizlib/-/minizlib-3.0.1.tgz",
+      "resolved": "https://registry.npmjs.org/minizlib/-/minizlib-3.1.0.tgz",
-      "integrity": "sha512-umcy022ILvb5/3Djuu8LWeqUa8D68JaBzlttKeMWen48SjabqS3iY5w/vzeMzMUNhLDifyhbOwKDSznB1vvrwg==",
+      "integrity": "sha512-KZxYo1BUkWD2TVFLr0MQoM8vUUigWD3LlD83a/75BqC+4qE0Hb1Vo5v1FgcfaNXvfXzr+5EhQ6ing/CaBijTlw==",
+      "license": "MIT",
       "dependencies": {
-        "minipass": "^7.0.4",
+        "minipass": "^7.1.2"
-        "rimraf": "^5.0.5"
       },
       "engines": {
         "node": ">= 18"
       }
     },
-    "node_modules/mkdirp": {
-      "version": "3.0.1",
-      "resolved": "https://registry.npmjs.org/mkdirp/-/mkdirp-3.0.1.tgz",
-      "integrity": "sha512-+NsyUUAZDmo6YVHzL/stxSu3t9YS1iljliy3BSDrXJ/dkn1KYdmtZODGGjLcc9XLgVVpH4KshHB8XmZgMhaBXg==",
-      "bin": {
-        "mkdirp": "dist/cjs/src/bin.js"
-      },
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
     "node_modules/ms": {
       "version": "2.1.3",
       "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz",
@@ -4010,11 +4034,6 @@
         "node": ">=6"
       }
     },
-    "node_modules/package-json-from-dist": {
-      "version": "1.0.1",
-      "resolved": "https://registry.npmjs.org/package-json-from-dist/-/package-json-from-dist-1.0.1.tgz",
-      "integrity": "sha512-UEZIS3/by4OC8vL3P2dTXRETpebLI2NiI5vIrjaD/5UtrkFX/tNbwjTSRAGC/+7CAo2pIcBaRgWmcBBHcsaCIw=="
-    },
     "node_modules/parse-json": {
       "version": "5.2.0",
       "resolved": "https://registry.npmjs.org/parse-json/-/parse-json-5.2.0.tgz",
@@ -4055,6 +4074,7 @@
       "version": "3.1.1",
       "resolved": "https://registry.npmjs.org/path-key/-/path-key-3.1.1.tgz",
       "integrity": "sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==",
+      "dev": true,
       "engines": {
         "node": ">=8"
       }
@@ -4065,26 +4085,6 @@
       "integrity": "sha512-LDJzPVEEEPR+y48z93A0Ed0yXb8pAByGWo/k5YYdYgpY2/2EsOsksJrq7lOHxryrVOn1ejG6oAp8ahvOIQD8sw==",
       "dev": true
     },
-    "node_modules/path-scurry": {
-      "version": "1.11.1",
-      "resolved": "https://registry.npmjs.org/path-scurry/-/path-scurry-1.11.1.tgz",
-      "integrity": "sha512-Xa4Nw17FS9ApQFJ9umLiJS4orGjm7ZzwUrwamcGQuHSzDyth9boKDaycYdDcZDuqYATXw4HFXgaqWTctW/v1HA==",
-      "dependencies": {
-        "lru-cache": "^10.2.0",
-        "minipass": "^5.0.0 || ^6.0.2 || ^7.0.0"
-      },
-      "engines": {
-        "node": ">=16 || 14 >=14.18"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/path-scurry/node_modules/lru-cache": {
-      "version": "10.4.3",
-      "resolved": "https://registry.npmjs.org/lru-cache/-/lru-cache-10.4.3.tgz",
-      "integrity": "sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ=="
-    },
     "node_modules/picocolors": {
       "version": "1.1.1",
       "resolved": "https://registry.npmjs.org/picocolors/-/picocolors-1.1.1.tgz",
@@ -4246,61 +4246,6 @@
         "node": ">=10"
       }
     },
-    "node_modules/rimraf": {
-      "version": "5.0.10",
-      "resolved": "https://registry.npmjs.org/rimraf/-/rimraf-5.0.10.tgz",
-      "integrity": "sha512-l0OE8wL34P4nJH/H2ffoaniAokM2qSmrtXHmlpvYr5AVVX8msAyW0l8NVJFDxlSK4u3Uh/f41cQheDVdnYijwQ==",
-      "dependencies": {
-        "glob": "^10.3.7"
-      },
-      "bin": {
-        "rimraf": "dist/esm/bin.mjs"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/rimraf/node_modules/brace-expansion": {
-      "version": "2.0.1",
-      "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz",
-      "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==",
-      "dependencies": {
-        "balanced-match": "^1.0.0"
-      }
-    },
-    "node_modules/rimraf/node_modules/glob": {
-      "version": "10.4.5",
-      "resolved": "https://registry.npmjs.org/glob/-/glob-10.4.5.tgz",
-      "integrity": "sha512-7Bv8RF0k6xjo7d4A/PxYLbUCfb6c+Vpd2/mB2yRDlew7Jb5hEXiCD9ibfO7wpk8i4sevK6DFny9h7EYbM3/sHg==",
-      "dependencies": {
-        "foreground-child": "^3.1.0",
-        "jackspeak": "^3.1.2",
-        "minimatch": "^9.0.4",
-        "minipass": "^7.1.2",
-        "package-json-from-dist": "^1.0.0",
-        "path-scurry": "^1.11.1"
-      },
-      "bin": {
-        "glob": "dist/esm/bin.mjs"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
-    "node_modules/rimraf/node_modules/minimatch": {
-      "version": "9.0.5",
-      "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.5.tgz",
-      "integrity": "sha512-G6T0ZX48xgozx7587koeX9Ys2NYy6Gmv//P89sEte9V9whIapMNF4idKxnW2QtCcLiTWlb/wfCabAtAFWhhBow==",
-      "dependencies": {
-        "brace-expansion": "^2.0.1"
-      },
-      "engines": {
-        "node": ">=16 || 14 >=14.17"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/isaacs"
-      }
-    },
     "node_modules/semver": {
       "version": "7.6.3",
       "resolved": "https://registry.npmjs.org/semver/-/semver-7.6.3.tgz",
@@ -4354,6 +4299,7 @@
       "version": "2.0.0",
       "resolved": "https://registry.npmjs.org/shebang-command/-/shebang-command-2.0.0.tgz",
       "integrity": "sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==",
+      "dev": true,
       "dependencies": {
         "shebang-regex": "^3.0.0"
       },
@@ -4365,6 +4311,7 @@
       "version": "3.0.0",
       "resolved": "https://registry.npmjs.org/shebang-regex/-/shebang-regex-3.0.0.tgz",
       "integrity": "sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==",
+      "dev": true,
       "engines": {
         "node": ">=8"
       }
@@ -4452,20 +4399,7 @@
       "version": "4.2.3",
       "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz",
       "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==",
-      "dependencies": {
+      "dev": true,
-        "emoji-regex": "^8.0.0",
-        "is-fullwidth-code-point": "^3.0.0",
-        "strip-ansi": "^6.0.1"
-      },
-      "engines": {
-        "node": ">=8"
-      }
-    },
-    "node_modules/string-width-cjs": {
-      "name": "string-width",
-      "version": "4.2.3",
-      "resolved": "https://registry.npmjs.org/string-width/-/string-width-4.2.3.tgz",
-      "integrity": "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==",
       "dependencies": {
         "emoji-regex": "^8.0.0",
         "is-fullwidth-code-point": "^3.0.0",
@@ -4479,18 +4413,7 @@
       "version": "6.0.1",
       "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz",
       "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==",
-      "dependencies": {
+      "dev": true,
-        "ansi-regex": "^5.0.1"
-      },
-      "engines": {
-        "node": ">=8"
-      }
-    },
-    "node_modules/strip-ansi-cjs": {
-      "name": "strip-ansi",
-      "version": "6.0.1",
-      "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-6.0.1.tgz",
-      "integrity": "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==",
       "dependencies": {
         "ansi-regex": "^5.0.1"
       },
@@ -4541,15 +4464,15 @@
       }
     },
     "node_modules/tar": {
-      "version": "7.4.3",
+      "version": "7.5.10",
-      "resolved": "https://registry.npmjs.org/tar/-/tar-7.4.3.tgz",
+      "resolved": "https://registry.npmjs.org/tar/-/tar-7.5.10.tgz",
-      "integrity": "sha512-5S7Va8hKfV7W5U6g3aYxXmlPoZVAwUMy9AOKyF2fVuZa2UD3qZjg578OrLRt8PcNN1PleVaL/5/yYATNL0ICUw==",
+      "integrity": "sha512-8mOPs1//5q/rlkNSPcCegA6hiHJYDmSLEI8aMH/CdSQJNWztHC9WHNam5zdQlfpTwB9Xp7IBEsHfV5LKMJGVAw==",
+      "license": "BlueOak-1.0.0",
       "dependencies": {
         "@isaacs/fs-minipass": "^4.0.0",
         "chownr": "^3.0.0",
         "minipass": "^7.1.2",
-        "minizlib": "^3.0.1",
+        "minizlib": "^3.1.0",
-        "mkdirp": "^3.0.1",
         "yallist": "^5.0.0"
       },
       "engines": {
@@ -4782,6 +4705,7 @@
       "version": "2.0.2",
       "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz",
       "integrity": "sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==",
+      "dev": true,
       "dependencies": {
         "isexe": "^2.0.0"
       },
@@ -4809,23 +4733,6 @@
         "url": "https://github.com/chalk/wrap-ansi?sponsor=1"
       }
     },
-    "node_modules/wrap-ansi-cjs": {
-      "name": "wrap-ansi",
-      "version": "7.0.0",
-      "resolved": "https://registry.npmjs.org/wrap-ansi/-/wrap-ansi-7.0.0.tgz",
-      "integrity": "sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==",
-      "dependencies": {
-        "ansi-styles": "^4.0.0",
-        "string-width": "^4.1.0",
-        "strip-ansi": "^6.0.0"
-      },
-      "engines": {
-        "node": ">=10"
-      },
-      "funding": {
-        "url": "https://github.com/chalk/wrap-ansi?sponsor=1"
-      }
-    },
     "node_modules/wrappy": {
       "version": "1.0.2",
       "resolved": "https://registry.npmjs.org/wrappy/-/wrappy-1.0.2.tgz",

nodejs/lancedb/arrow.ts

@@ -20,6 +20,8 @@ import {
   Float32,
   Float64,
   Int,
+  Int8,
+  Int16,
   Int32,
   Int64,
   LargeBinary,
@@ -35,6 +37,8 @@ import {
   Timestamp,
   Type,
   Uint8,
+  Uint16,
+  Uint32,
   Utf8,
   Vector,
   makeVector as arrowMakeVector,
@@ -113,8 +117,9 @@ export type TableLike =
 export type IntoVector =
   | Float32Array
   | Float64Array
+  | Uint8Array
   | number[]
-  | Promise<Float32Array | Float64Array | number[]>;
+  | Promise<Float32Array | Float64Array | Uint8Array | number[]>;
 
 export type MultiVector = IntoVector[];
 
@@ -122,14 +127,48 @@ export function isMultiVector(value: unknown): value is MultiVector {
   return Array.isArray(value) && isIntoVector(value[0]);
 }
 
+// Float16Array is not in TypeScript's standard lib yet; access dynamically
+type Float16ArrayCtor = new (
+  ...args: unknown[]
+) => { buffer: ArrayBuffer; byteOffset: number; byteLength: number };
+const float16ArrayCtor = (globalThis as unknown as Record<string, unknown>)
+  .Float16Array as Float16ArrayCtor | undefined;
+
 export function isIntoVector(value: unknown): value is IntoVector {
   return (
     value instanceof Float32Array ||
     value instanceof Float64Array ||
+    value instanceof Uint8Array ||
+    (float16ArrayCtor !== undefined && value instanceof float16ArrayCtor) ||
     (Array.isArray(value) && !Array.isArray(value[0]))
   );
 }
 
+/**
+ * Extract the underlying byte buffer and data type from a typed array
+ * for passing to the Rust NAPI layer without precision loss.
+ */
+export function extractVectorBuffer(
+  vector: Float32Array | Float64Array | Uint8Array,
+): { data: Uint8Array; dtype: string } | null {
+  if (float16ArrayCtor !== undefined && vector instanceof float16ArrayCtor) {
+    return {
+      data: new Uint8Array(vector.buffer, vector.byteOffset, vector.byteLength),
+      dtype: "float16",
+    };
+  }
+  if (vector instanceof Float64Array) {
+    return {
+      data: new Uint8Array(vector.buffer, vector.byteOffset, vector.byteLength),
+      dtype: "float64",
+    };
+  }
+  if (vector instanceof Uint8Array && !(vector instanceof Float32Array)) {
+    return { data: vector, dtype: "uint8" };
+  }
+  return null;
+}
+
 export function isArrowTable(value: object): value is TableLike {
   if (value instanceof ArrowTable) return true;
   return "schema" in value && "batches" in value;
@@ -529,7 +568,8 @@ function isObject(value: unknown): value is Record<string, unknown> {
     !(value instanceof Date) &&
     !(value instanceof Set) &&
     !(value instanceof Map) &&
-    !(value instanceof Buffer)
+    !(value instanceof Buffer) &&
+    !ArrayBuffer.isView(value)
   );
 }
 
@@ -588,6 +628,13 @@ function inferType(
     return new Bool();
   } else if (value instanceof Buffer) {
     return new Binary();
+  } else if (ArrayBuffer.isView(value) && !(value instanceof DataView)) {
+    const info = typedArrayToArrowType(value);
+    if (info !== undefined) {
+      const child = new Field("item", info.elementType, true);
+      return new FixedSizeList(info.length, child);
+    }
+    return undefined;
   } else if (Array.isArray(value)) {
     if (value.length === 0) {
       return undefined; // Without any values we can't infer the type
@@ -746,6 +793,32 @@ function makeListVector(lists: unknown[][]): Vector<unknown> {
   return listBuilder.finish().toVector();
 }
 
+/**
+ * Map a JS TypedArray instance to the corresponding Arrow element DataType
+ * and its length. Returns undefined if the value is not a recognized TypedArray.
+ */
+function typedArrayToArrowType(
+  value: ArrayBufferView,
+): { elementType: DataType; length: number } | undefined {
+  if (value instanceof Float32Array)
+    return { elementType: new Float32(), length: value.length };
+  if (value instanceof Float64Array)
+    return { elementType: new Float64(), length: value.length };
+  if (value instanceof Uint8Array)
+    return { elementType: new Uint8(), length: value.length };
+  if (value instanceof Uint16Array)
+    return { elementType: new Uint16(), length: value.length };
+  if (value instanceof Uint32Array)
+    return { elementType: new Uint32(), length: value.length };
+  if (value instanceof Int8Array)
+    return { elementType: new Int8(), length: value.length };
+  if (value instanceof Int16Array)
+    return { elementType: new Int16(), length: value.length };
+  if (value instanceof Int32Array)
+    return { elementType: new Int32(), length: value.length };
+  return undefined;
+}
+
 /** Helper function to convert an Array of JS values to an Arrow Vector */
 function makeVector(
   values: unknown[],
@@ -814,6 +887,16 @@ function makeVector(
       "makeVector cannot infer the type if all values are null or undefined",
     );
   }
+  if (ArrayBuffer.isView(sampleValue) && !(sampleValue instanceof DataView)) {
+    const info = typedArrayToArrowType(sampleValue);
+    if (info !== undefined) {
+      const fslType = new FixedSizeList(
+        info.length,
+        new Field("item", info.elementType, true),
+      );
+      return vectorFromArray(values, fslType);
+    }
+  }
   if (Array.isArray(sampleValue)) {
     // Default Arrow inference doesn't handle list types
     return makeListVector(values as unknown[][]);

nodejs/lancedb/connection.ts

@@ -166,25 +166,25 @@ export abstract class Connection {
    * List all the table names in this database.
    *
    * Tables will be returned in lexicographical order.
-   * @param {string[]} namespace - The namespace to list tables from (defaults to root namespace)
+   * @param {string[]} namespacePath - The namespace path to list tables from (defaults to root namespace)
    * @param {Partial<TableNamesOptions>} options - options to control the
    * paging / start point
    *
    */
   abstract tableNames(
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<TableNamesOptions>,
   ): Promise<string[]>;
 
   /**
    * Open a table in the database.
    * @param {string} name - The name of the table
-   * @param {string[]} namespace - The namespace of the table (defaults to root namespace)
+   * @param {string[]} namespacePath - The namespace path of the table (defaults to root namespace)
    * @param {Partial<OpenTableOptions>} options - Additional options
    */
   abstract openTable(
     name: string,
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<OpenTableOptions>,
   ): Promise<Table>;
 
@@ -193,7 +193,7 @@ export abstract class Connection {
    * @param {object} options - The options object.
    * @param {string} options.name - The name of the table.
    * @param {Data} options.data - Non-empty Array of Records to be inserted into the table
-   * @param {string[]} namespace - The namespace to create the table in (defaults to root namespace)
+   * @param {string[]} namespacePath - The namespace path to create the table in (defaults to root namespace)
    *
    */
   abstract createTable(
@@ -201,7 +201,7 @@ export abstract class Connection {
       name: string;
       data: Data;
     } & Partial<CreateTableOptions>,
-    namespace?: string[],
+    namespacePath?: string[],
   ): Promise<Table>;
   /**
    * Creates a new Table and initialize it with new data.
@@ -220,13 +220,13 @@ export abstract class Connection {
    * @param {string} name - The name of the table.
    * @param {Record<string, unknown>[] | TableLike} data - Non-empty Array of Records
    * to be inserted into the table
-   * @param {string[]} namespace - The namespace to create the table in (defaults to root namespace)
+   * @param {string[]} namespacePath - The namespace path to create the table in (defaults to root namespace)
    * @param {Partial<CreateTableOptions>} options - Additional options
    */
   abstract createTable(
     name: string,
     data: Record<string, unknown>[] | TableLike,
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<CreateTableOptions>,
   ): Promise<Table>;
 
@@ -245,28 +245,28 @@ export abstract class Connection {
    * Creates a new empty Table
    * @param {string} name - The name of the table.
    * @param {Schema} schema - The schema of the table
-   * @param {string[]} namespace - The namespace to create the table in (defaults to root namespace)
+   * @param {string[]} namespacePath - The namespace path to create the table in (defaults to root namespace)
    * @param {Partial<CreateTableOptions>} options - Additional options
    */
   abstract createEmptyTable(
     name: string,
     schema: import("./arrow").SchemaLike,
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<CreateTableOptions>,
   ): Promise<Table>;
 
   /**
    * Drop an existing table.
    * @param {string} name The name of the table to drop.
-   * @param {string[]} namespace The namespace of the table (defaults to root namespace).
+   * @param {string[]} namespacePath The namespace path of the table (defaults to root namespace).
    */
-  abstract dropTable(name: string, namespace?: string[]): Promise<void>;
+  abstract dropTable(name: string, namespacePath?: string[]): Promise<void>;
 
   /**
    * Drop all tables in the database.
-   * @param {string[]} namespace The namespace to drop tables from (defaults to root namespace).
+   * @param {string[]} namespacePath The namespace path to drop tables from (defaults to root namespace).
    */
-  abstract dropAllTables(namespace?: string[]): Promise<void>;
+  abstract dropAllTables(namespacePath?: string[]): Promise<void>;
 
   /**
    * Clone a table from a source table.
@@ -279,7 +279,7 @@ export abstract class Connection {
    * @param {string} targetTableName - The name of the target table to create.
    * @param {string} sourceUri - The URI of the source table to clone from.
    * @param {object} options - Clone options.
-   * @param {string[]} options.targetNamespace - The namespace for the target table (defaults to root namespace).
+   * @param {string[]} options.targetNamespacePath - The namespace path for the target table (defaults to root namespace).
    * @param {number} options.sourceVersion - The version of the source table to clone.
    * @param {string} options.sourceTag - The tag of the source table to clone.
    * @param {boolean} options.isShallow - Whether to perform a shallow clone (defaults to true).
@@ -288,7 +288,7 @@ export abstract class Connection {
     targetTableName: string,
     sourceUri: string,
     options?: {
-      targetNamespace?: string[];
+      targetNamespacePath?: string[];
       sourceVersion?: number;
       sourceTag?: string;
       isShallow?: boolean;
@@ -319,25 +319,25 @@ export class LocalConnection extends Connection {
   }
 
   async tableNames(
-    namespaceOrOptions?: string[] | Partial<TableNamesOptions>,
+    namespacePathOrOptions?: string[] | Partial<TableNamesOptions>,
     options?: Partial<TableNamesOptions>,
   ): Promise<string[]> {
-    // Detect if first argument is namespace array or options object
+    // Detect if first argument is namespacePath array or options object
-    let namespace: string[] | undefined;
+    let namespacePath: string[] | undefined;
     let tableNamesOptions: Partial<TableNamesOptions> | undefined;
 
-    if (Array.isArray(namespaceOrOptions)) {
+    if (Array.isArray(namespacePathOrOptions)) {
-      // First argument is namespace array
+      // First argument is namespacePath array
-      namespace = namespaceOrOptions;
+      namespacePath = namespacePathOrOptions;
       tableNamesOptions = options;
     } else {
       // First argument is options object (backwards compatibility)
-      namespace = undefined;
+      namespacePath = undefined;
-      tableNamesOptions = namespaceOrOptions;
+      tableNamesOptions = namespacePathOrOptions;
     }
 
     return this.inner.tableNames(
-      namespace ?? [],
+      namespacePath ?? [],
       tableNamesOptions?.startAfter,
       tableNamesOptions?.limit,
     );
@@ -345,12 +345,12 @@ export class LocalConnection extends Connection {
 
   async openTable(
     name: string,
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<OpenTableOptions>,
   ): Promise<Table> {
     const innerTable = await this.inner.openTable(
       name,
-      namespace ?? [],
+      namespacePath ?? [],
       cleanseStorageOptions(options?.storageOptions),
       options?.indexCacheSize,
     );
@@ -362,7 +362,7 @@ export class LocalConnection extends Connection {
     targetTableName: string,
     sourceUri: string,
     options?: {
-      targetNamespace?: string[];
+      targetNamespacePath?: string[];
       sourceVersion?: number;
       sourceTag?: string;
       isShallow?: boolean;
@@ -371,7 +371,7 @@ export class LocalConnection extends Connection {
     const innerTable = await this.inner.cloneTable(
       targetTableName,
       sourceUri,
-      options?.targetNamespace ?? [],
+      options?.targetNamespacePath ?? [],
       options?.sourceVersion ?? null,
       options?.sourceTag ?? null,
       options?.isShallow ?? true,
@@ -406,42 +406,42 @@ export class LocalConnection extends Connection {
     nameOrOptions:
       | string
       | ({ name: string; data: Data } & Partial<CreateTableOptions>),
-    dataOrNamespace?: Record<string, unknown>[] | TableLike | string[],
+    dataOrNamespacePath?: Record<string, unknown>[] | TableLike | string[],
-    namespaceOrOptions?: string[] | Partial<CreateTableOptions>,
+    namespacePathOrOptions?: string[] | Partial<CreateTableOptions>,
     options?: Partial<CreateTableOptions>,
   ): Promise<Table> {
     if (typeof nameOrOptions !== "string" && "name" in nameOrOptions) {
-      // First overload: createTable(options, namespace?)
+      // First overload: createTable(options, namespacePath?)
       const { name, data, ...createOptions } = nameOrOptions;
-      const namespace = dataOrNamespace as string[] | undefined;
+      const namespacePath = dataOrNamespacePath as string[] | undefined;
-      return this._createTableImpl(name, data, namespace, createOptions);
+      return this._createTableImpl(name, data, namespacePath, createOptions);
     }
 
-    // Second overload: createTable(name, data, namespace?, options?)
+    // Second overload: createTable(name, data, namespacePath?, options?)
     const name = nameOrOptions;
-    const data = dataOrNamespace as Record<string, unknown>[] | TableLike;
+    const data = dataOrNamespacePath as Record<string, unknown>[] | TableLike;
 
-    // Detect if third argument is namespace array or options object
+    // Detect if third argument is namespacePath array or options object
-    let namespace: string[] | undefined;
+    let namespacePath: string[] | undefined;
     let createOptions: Partial<CreateTableOptions> | undefined;
 
-    if (Array.isArray(namespaceOrOptions)) {
+    if (Array.isArray(namespacePathOrOptions)) {
-      // Third argument is namespace array
+      // Third argument is namespacePath array
-      namespace = namespaceOrOptions;
+      namespacePath = namespacePathOrOptions;
       createOptions = options;
     } else {
       // Third argument is options object (backwards compatibility)
-      namespace = undefined;
+      namespacePath = undefined;
-      createOptions = namespaceOrOptions;
+      createOptions = namespacePathOrOptions;
     }
 
-    return this._createTableImpl(name, data, namespace, createOptions);
+    return this._createTableImpl(name, data, namespacePath, createOptions);
   }
 
   private async _createTableImpl(
     name: string,
     data: Data,
-    namespace?: string[],
+    namespacePath?: string[],
     options?: Partial<CreateTableOptions>,
   ): Promise<Table> {
     if (data === undefined) {
@@ -455,7 +455,7 @@ export class LocalConnection extends Connection {
       name,
       buf,
       mode,
-      namespace ?? [],
+      namespacePath ?? [],
       storageOptions,
     );
 
@@ -465,21 +465,21 @@ export class LocalConnection extends Connection {
   async createEmptyTable(
     name: string,
     schema: import("./arrow").SchemaLike,
-    namespaceOrOptions?: string[] | Partial<CreateTableOptions>,
+    namespacePathOrOptions?: string[] | Partial<CreateTableOptions>,
     options?: Partial<CreateTableOptions>,
   ): Promise<Table> {
-    // Detect if third argument is namespace array or options object
+    // Detect if third argument is namespacePath array or options object
-    let namespace: string[] | undefined;
+    let namespacePath: string[] | undefined;
     let createOptions: Partial<CreateTableOptions> | undefined;
 
-    if (Array.isArray(namespaceOrOptions)) {
+    if (Array.isArray(namespacePathOrOptions)) {
-      // Third argument is namespace array
+      // Third argument is namespacePath array
-      namespace = namespaceOrOptions;
+      namespacePath = namespacePathOrOptions;
       createOptions = options;
     } else {
       // Third argument is options object (backwards compatibility)
-      namespace = undefined;
+      namespacePath = undefined;
-      createOptions = namespaceOrOptions;
+      createOptions = namespacePathOrOptions;
     }
 
     let mode: string = createOptions?.mode ?? "create";
@@ -502,18 +502,18 @@ export class LocalConnection extends Connection {
       name,
       buf,
       mode,
-      namespace ?? [],
+      namespacePath ?? [],
       storageOptions,
     );
     return new LocalTable(innerTable);
   }
 
-  async dropTable(name: string, namespace?: string[]): Promise<void> {
+  async dropTable(name: string, namespacePath?: string[]): Promise<void> {
-    return this.inner.dropTable(name, namespace ?? []);
+    return this.inner.dropTable(name, namespacePath ?? []);
   }
 
-  async dropAllTables(namespace?: string[]): Promise<void> {
+  async dropAllTables(namespacePath?: string[]): Promise<void> {
-    return this.inner.dropAllTables(namespace ?? []);
+    return this.inner.dropAllTables(namespacePath ?? []);
   }
 }
 

nodejs/lancedb/index.ts

@@ -273,7 +273,9 @@ export async function connect(
   let nativeProvider: NativeJsHeaderProvider | undefined;
   if (finalHeaderProvider) {
     if (typeof finalHeaderProvider === "function") {
-      nativeProvider = new NativeJsHeaderProvider(finalHeaderProvider);
+      nativeProvider = new NativeJsHeaderProvider(async () =>
+        finalHeaderProvider(),
+      );
     } else if (
       finalHeaderProvider &&
       typeof finalHeaderProvider.getHeaders === "function"

nodejs/lancedb/query.ts

@@ -5,6 +5,7 @@ import {
   Table as ArrowTable,
   type IntoVector,
   RecordBatch,
+  extractVectorBuffer,
   fromBufferToRecordBatch,
   fromRecordBatchToBuffer,
   tableFromIPC,
@@ -661,10 +662,8 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
       const res = (async () => {
         try {
           const v = await vector;
-          const arr = Float32Array.from(v);
-          //
           // biome-ignore lint/suspicious/noExplicitAny: we need to get the `inner`, but js has no package scoping
-          const value: any = this.addQueryVector(arr);
+          const value: any = this.addQueryVector(v);
           const inner = value.inner as
             | NativeVectorQuery
             | Promise<NativeVectorQuery>;
@@ -676,7 +675,12 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
       return new VectorQuery(res);
     } else {
       super.doCall((inner) => {
-        inner.addQueryVector(Float32Array.from(vector));
+        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;
     }
@@ -684,19 +688,17 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
 
   rerank(reranker: Reranker): VectorQuery {
     super.doCall((inner) =>
-      inner.rerank({
+      inner.rerank(async (args) => {
-        rerankHybrid: async (_, args) => {
+        const vecResults = await fromBufferToRecordBatch(args.vecResults);
-          const vecResults = await fromBufferToRecordBatch(args.vecResults);
+        const ftsResults = await fromBufferToRecordBatch(args.ftsResults);
-          const ftsResults = await fromBufferToRecordBatch(args.ftsResults);
+        const result = await reranker.rerankHybrid(
-          const result = await reranker.rerankHybrid(
+          args.query,
-            args.query,
+          vecResults as RecordBatch,
-            vecResults as RecordBatch,
+          ftsResults as RecordBatch,
-            ftsResults as RecordBatch,
+        );
-          );
 
-          const buffer = fromRecordBatchToBuffer(result);
+        const buffer = fromRecordBatchToBuffer(result);
-          return buffer;
+        return buffer;
-        },
       }),
     );
 
@@ -767,14 +769,23 @@ export class Query extends StandardQueryBase<NativeQuery> {
    * a default `limit` of 10 will be used.  @see {@link Query#limit}
    */
   nearestTo(vector: IntoVector): VectorQuery {
+    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) => {
-        if (vector instanceof Promise) {
+        const resolved = vector instanceof Promise ? await vector : vector;
-          const arr = await vector.then((v) => Float32Array.from(v));
+        return callNearestTo(inner, resolved);
-          return inner.nearestTo(arr);
-        } else {
-          return inner.nearestTo(Float32Array.from(vector));
-        }
       });
       return new VectorQuery(nativeQuery);
     }
@@ -782,10 +793,8 @@ export class Query extends StandardQueryBase<NativeQuery> {
       const res = (async () => {
         try {
           const v = await vector;
-          const arr = Float32Array.from(v);
-          //
           // biome-ignore lint/suspicious/noExplicitAny: we need to get the `inner`, but js has no package scoping
-          const value: any = this.nearestTo(arr);
+          const value: any = this.nearestTo(v);
           const inner = value.inner as
             | NativeVectorQuery
             | Promise<NativeVectorQuery>;
@@ -796,7 +805,7 @@ export class Query extends StandardQueryBase<NativeQuery> {
       })();
       return new VectorQuery(res);
     } else {
-      const vectorQuery = this.inner.nearestTo(Float32Array.from(vector));
+      const vectorQuery = callNearestTo(this.inner, vector);
       return new VectorQuery(vectorQuery);
     }
   }

nodejs/lancedb/table.ts

@@ -5,12 +5,15 @@ import {
   Table as ArrowTable,
   Data,
   DataType,
+  Field,
   IntoVector,
   MultiVector,
   Schema,
   dataTypeToJson,
   fromDataToBuffer,
+  fromTableToBuffer,
   isMultiVector,
+  makeEmptyTable,
   tableFromIPC,
 } from "./arrow";
 
@@ -84,6 +87,16 @@ export interface OptimizeOptions {
    * tbl.optimize({cleanupOlderThan: new Date()});
    */
   cleanupOlderThan: Date;
+  /**
+   * Because they may be part of an in-progress transaction, files newer than
+   * 7 days old are not deleted by default. If you are sure that there are no
+   * in-progress transactions, then you can set this to true to delete all
+   * files older than `cleanupOlderThan`.
+   *
+   * **WARNING**: This should only be set to true if you can guarantee that
+   * no other process is currently working on this dataset. Otherwise the
+   * dataset could be put into a corrupted state.
+   */
   deleteUnverified: boolean;
 }
 
@@ -381,15 +394,16 @@ export abstract class Table {
   abstract vectorSearch(vector: IntoVector | MultiVector): VectorQuery;
   /**
    * Add new columns with defined values.
-   * @param {AddColumnsSql[]} newColumnTransforms pairs of column names and
+   * @param {AddColumnsSql[] | Field | Field[] | Schema} newColumnTransforms Either:
-   * the SQL expression to use to calculate the value of the new column. These
+   *   - An array of objects with column names and SQL expressions to calculate values
-   * expressions will be evaluated for each row in the table, and can
+   *   - A single Arrow Field defining one column with its data type (column will be initialized with null values)
-   * reference existing columns in the table.
+   *   - An array of Arrow Fields defining columns with their data types (columns will be initialized with null values)
+   *   - An Arrow Schema defining columns with their data types (columns will be initialized with null values)
    * @returns {Promise<AddColumnsResult>} A promise that resolves to an object
    * containing the new version number of the table after adding the columns.
    */
   abstract addColumns(
-    newColumnTransforms: AddColumnsSql[],
+    newColumnTransforms: AddColumnsSql[] | Field | Field[] | Schema,
   ): Promise<AddColumnsResult>;
 
   /**
@@ -501,19 +515,7 @@ export abstract class Table {
    *  - Index: Optimizes the indices, adding new data to existing indices
    *
    *
-   *  Experimental API
+   *  The frequency an application should call optimize is based on the frequency of
-   *  ----------------
-   *
-   *  The optimization process is undergoing active development and may change.
-   *  Our goal with these changes is to improve the performance of optimization and
-   *  reduce the complexity.
-   *
-   *  That being said, it is essential today to run optimize if you want the best
-   *  performance.  It should be stable and safe to use in production, but it our
-   *  hope that the API may be simplified (or not even need to be called) in the
-   *  future.
-   *
-   *  The frequency an application shoudl call optimize is based on the frequency of
    *  data modifications.  If data is frequently added, deleted, or updated then
    *  optimize should be run frequently.  A good rule of thumb is to run optimize if
    *  you have added or modified 100,000 or more records or run more than 20 data
@@ -806,9 +808,40 @@ export class LocalTable extends Table {
   // TODO: Support BatchUDF
 
   async addColumns(
-    newColumnTransforms: AddColumnsSql[],
+    newColumnTransforms: AddColumnsSql[] | Field | Field[] | Schema,
   ): Promise<AddColumnsResult> {
-    return await this.inner.addColumns(newColumnTransforms);
+    // Handle single Field -> convert to array of Fields
+    if (newColumnTransforms instanceof Field) {
+      newColumnTransforms = [newColumnTransforms];
+    }
+
+    // Handle array of Fields -> convert to Schema
+    if (
+      Array.isArray(newColumnTransforms) &&
+      newColumnTransforms.length > 0 &&
+      newColumnTransforms[0] instanceof Field
+    ) {
+      const fields = newColumnTransforms as Field[];
+      newColumnTransforms = new Schema(fields);
+    }
+
+    // Handle Schema -> use schema-based approach
+    if (newColumnTransforms instanceof Schema) {
+      const schema = newColumnTransforms;
+      // Convert schema to buffer using Arrow IPC format
+      const emptyTable = makeEmptyTable(schema);
+      const schemaBuf = await fromTableToBuffer(emptyTable);
+      return await this.inner.addColumnsWithSchema(schemaBuf);
+    }
+
+    // Handle SQL expressions (existing functionality)
+    if (Array.isArray(newColumnTransforms)) {
+      return await this.inner.addColumns(
+        newColumnTransforms as AddColumnsSql[],
+      );
+    }
+
+    throw new Error("Invalid input type for addColumns");
   }
 
   async alterColumns(

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

nodejs/npm/linux-arm64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

nodejs/npm/linux-arm64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

nodejs/npm/linux-x64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

nodejs/npm/linux-x64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

nodejs/npm/win32-arm64-msvc/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.27.0-beta.0",
+  "version": "0.27.2",
   "os": [
     "win32"
   ],

nodejs/npm/win32-x64-msvc/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.27.0-beta.0",
+	"version": "0.27.2",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.27.0-beta.0",
+  "version": "0.27.2",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",
@@ -21,19 +21,16 @@
   },
   "types": "dist/index.d.ts",
   "napi": {
-    "name": "lancedb",
+    "binaryName": "lancedb",
-    "triples": {
+    "targets": [
-      "defaults": false,
+      "aarch64-apple-darwin",
-      "additional": [
+      "x86_64-unknown-linux-gnu",
-        "aarch64-apple-darwin",
+      "aarch64-unknown-linux-gnu",
-        "x86_64-unknown-linux-gnu",
+      "x86_64-unknown-linux-musl",
-        "aarch64-unknown-linux-gnu",
+      "aarch64-unknown-linux-musl",
-        "x86_64-unknown-linux-musl",
+      "x86_64-pc-windows-msvc",
-        "aarch64-unknown-linux-musl",
+      "aarch64-pc-windows-msvc"
-        "x86_64-pc-windows-msvc",
+    ]
-        "aarch64-pc-windows-msvc"
-      ]
-    }
   },
   "license": "Apache-2.0",
   "repository": {
@@ -46,7 +43,7 @@
     "@aws-sdk/client-s3": "^3.33.0",
     "@biomejs/biome": "^1.7.3",
     "@jest/globals": "^29.7.0",
-    "@napi-rs/cli": "^2.18.3",
+    "@napi-rs/cli": "^3.5.1",
     "@types/axios": "^0.14.0",
     "@types/jest": "^29.1.2",
     "@types/node": "^22.7.4",
@@ -75,9 +72,9 @@
   "os": ["darwin", "linux", "win32"],
   "scripts": {
     "artifacts": "napi artifacts",
-    "build:debug": "napi build --platform --no-const-enum --dts ../lancedb/native.d.ts --js ../lancedb/native.js lancedb",
+    "build:debug": "napi build --platform --dts ../lancedb/native.d.ts --js ../lancedb/native.js --output-dir lancedb",
     "postbuild:debug": "shx mkdir -p dist && shx cp lancedb/*.node dist/",
-    "build:release": "napi build --platform --no-const-enum --release --dts ../lancedb/native.d.ts --js ../lancedb/native.js dist/",
+    "build:release": "napi build --platform --release --dts ../lancedb/native.d.ts --js ../lancedb/native.js --output-dir dist",
     "postbuild:release": "shx mkdir -p dist && shx cp lancedb/*.node dist/",
     "build": "npm run build:debug && npm run tsc",
     "build-release": "npm run build:release && npm run tsc",
@@ -91,7 +88,7 @@
     "prepublishOnly": "napi prepublish -t npm",
     "test": "jest --verbose",
     "integration": "S3_TEST=1 npm run test",
-    "universal": "napi universal",
+    "universal": "napi universalize",
     "version": "napi version"
   },
   "dependencies": {

nodejs/src/connection.rs

@@ -8,10 +8,10 @@ use lancedb::database::{CreateTableMode, Database};
 use napi::bindgen_prelude::*;
 use napi_derive::*;
 
+use crate::ConnectionOptions;
 use crate::error::NapiErrorExt;
 use crate::header::JsHeaderProvider;
 use crate::table::Table;
-use crate::ConnectionOptions;
 use lancedb::connection::{ConnectBuilder, Connection as LanceDBConnection};
 
 use lancedb::ipc::{ipc_file_to_batches, ipc_file_to_schema};
@@ -119,12 +119,12 @@ impl Connection {
     #[napi(catch_unwind)]
     pub async fn table_names(
         &self,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         start_after: Option<String>,
         limit: Option<u32>,
     ) -> napi::Result<Vec<String>> {
         let mut op = self.get_inner()?.table_names();
-        op = op.namespace(namespace);
+        op = op.namespace(namespace_path.unwrap_or_default());
         if let Some(start_after) = start_after {
             op = op.start_after(start_after);
         }
@@ -146,7 +146,7 @@ impl Connection {
         name: String,
         buf: Buffer,
         mode: String,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
     ) -> napi::Result<Table> {
         let batches = ipc_file_to_batches(buf.to_vec())
@@ -154,7 +154,7 @@ impl Connection {
         let mode = Self::parse_create_mode_str(&mode)?;
         let mut builder = self.get_inner()?.create_table(&name, batches).mode(mode);
 
-        builder = builder.namespace(namespace);
+        builder = builder.namespace(namespace_path.unwrap_or_default());
 
         if let Some(storage_options) = storage_options {
             for (key, value) in storage_options {
@@ -171,7 +171,7 @@ impl Connection {
         name: String,
         schema_buf: Buffer,
         mode: String,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
     ) -> napi::Result<Table> {
         let schema = ipc_file_to_schema(schema_buf.to_vec()).map_err(|e| {
@@ -183,7 +183,7 @@ impl Connection {
             .create_empty_table(&name, schema)
             .mode(mode);
 
-        builder = builder.namespace(namespace);
+        builder = builder.namespace(namespace_path.unwrap_or_default());
 
         if let Some(storage_options) = storage_options {
             for (key, value) in storage_options {
@@ -198,13 +198,13 @@ impl Connection {
     pub async fn open_table(
         &self,
         name: String,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
         index_cache_size: Option<u32>,
     ) -> napi::Result<Table> {
         let mut builder = self.get_inner()?.open_table(&name);
 
-        builder = builder.namespace(namespace);
+        builder = builder.namespace(namespace_path.unwrap_or_default());
 
         if let Some(storage_options) = storage_options {
             for (key, value) in storage_options {
@@ -223,7 +223,7 @@ impl Connection {
         &self,
         target_table_name: String,
         source_uri: String,
-        target_namespace: Vec<String>,
+        target_namespace_path: Option<Vec<String>>,
         source_version: Option<i64>,
         source_tag: Option<String>,
         is_shallow: bool,
@@ -232,7 +232,7 @@ impl Connection {
             .get_inner()?
             .clone_table(&target_table_name, &source_uri);
 
-        builder = builder.target_namespace(target_namespace);
+        builder = builder.target_namespace(target_namespace_path.unwrap_or_default());
 
         if let Some(version) = source_version {
             builder = builder.source_version(version as u64);
@@ -250,18 +250,21 @@ impl Connection {
 
     /// Drop table with the name. Or raise an error if the table does not exist.
     #[napi(catch_unwind)]
-    pub async fn drop_table(&self, name: String, namespace: Vec<String>) -> napi::Result<()> {
+    pub async fn drop_table(
+        &self,
+        name: String,
+        namespace_path: Option<Vec<String>>,
+    ) -> napi::Result<()> {
+        let ns = namespace_path.unwrap_or_default();
         self.get_inner()?
-            .drop_table(&name, &namespace)
+            .drop_table(&name, &ns)
             .await
             .default_error()
     }
 
     #[napi(catch_unwind)]
-    pub async fn drop_all_tables(&self, namespace: Vec<String>) -> napi::Result<()> {
+    pub async fn drop_all_tables(&self, namespace_path: Option<Vec<String>>) -> napi::Result<()> {
-        self.get_inner()?
+        let ns = namespace_path.unwrap_or_default();
-            .drop_all_tables(&namespace)
+        self.get_inner()?.drop_all_tables(&ns).await.default_error()
-            .await
-            .default_error()
     }
 }

nodejs/src/header.rs

@@ -1,20 +1,19 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
-use napi::{
+use napi::{bindgen_prelude::*, threadsafe_function::ThreadsafeFunction};
-    bindgen_prelude::*,
-    threadsafe_function::{ErrorStrategy, ThreadsafeFunction},
-};
 use napi_derive::napi;
 use std::collections::HashMap;
 use std::sync::Arc;
 
+type GetHeadersFn = ThreadsafeFunction<(), Promise<HashMap<String, String>>, (), Status, false>;
+
 /// JavaScript HeaderProvider implementation that wraps a JavaScript callback.
 /// This is the only native header provider - all header provider implementations
 /// should provide a JavaScript function that returns headers.
 #[napi]
 pub struct JsHeaderProvider {
-    get_headers_fn: Arc<ThreadsafeFunction<(), ErrorStrategy::CalleeHandled>>,
+    get_headers_fn: Arc<GetHeadersFn>,
 }
 
 impl Clone for JsHeaderProvider {
@@ -29,9 +28,12 @@ impl Clone for JsHeaderProvider {
 impl JsHeaderProvider {
     /// Create a new JsHeaderProvider from a JavaScript callback
     #[napi(constructor)]
-    pub fn new(get_headers_callback: JsFunction) -> Result<Self> {
+    pub fn new(
+        get_headers_callback: Function<(), Promise<HashMap<String, String>>>,
+    ) -> Result<Self> {
         let get_headers_fn = get_headers_callback
-            .create_threadsafe_function(0, |ctx| Ok(vec![ctx.value]))
+            .build_threadsafe_function()
+            .build()
             .map_err(|e| {
                 Error::new(
                     Status::GenericFailure,
@@ -51,7 +53,7 @@ impl lancedb::remote::HeaderProvider for JsHeaderProvider {
     async fn get_headers(&self) -> lancedb::error::Result<HashMap<String, String>> {
         // Call the JavaScript function asynchronously
         let promise: Promise<HashMap<String, String>> =
-            self.get_headers_fn.call_async(Ok(())).await.map_err(|e| {
+            self.get_headers_fn.call_async(()).await.map_err(|e| {
                 lancedb::error::Error::Runtime {
                     message: format!("Failed to call JavaScript get_headers: {}", e),
                 }

nodejs/src/index.rs

@@ -3,12 +3,12 @@
 
 use std::sync::Mutex;
 
+use lancedb::index::Index as LanceDbIndex;
 use lancedb::index::scalar::{BTreeIndexBuilder, FtsIndexBuilder};
 use lancedb::index::vector::{
     IvfFlatIndexBuilder, IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder,
     IvfRqIndexBuilder,
 };
-use lancedb::index::Index as LanceDbIndex;
 use napi_derive::napi;
 
 use crate::util::parse_distance_type;

nodejs/src/lib.rs

@@ -60,7 +60,7 @@ pub struct OpenTableOptions {
     pub storage_options: Option<HashMap<String, String>>,
 }
 
-#[napi::module_init]
+#[napi_derive::module_init]
 fn init() {
     let env = Env::new()
         .filter_or("LANCEDB_LOG", "warn")

nodejs/src/query.rs

@@ -3,6 +3,12 @@
 
 use std::sync::Arc;
 
+use arrow_array::{
+    Array, Float16Array as ArrowFloat16Array, Float32Array as ArrowFloat32Array,
+    Float64Array as ArrowFloat64Array, UInt8Array as ArrowUInt8Array,
+};
+use arrow_buffer::ScalarBuffer;
+use half::f16;
 use lancedb::index::scalar::{
     BooleanQuery, BoostQuery, FtsQuery, FullTextSearchQuery, MatchQuery, MultiMatchQuery, Occur,
     Operator, PhraseQuery,
@@ -17,13 +23,40 @@ use lancedb::query::VectorQuery as LanceDbVectorQuery;
 use napi::bindgen_prelude::*;
 use napi_derive::napi;
 
-use crate::error::convert_error;
 use crate::error::NapiErrorExt;
+use crate::error::convert_error;
 use crate::iterator::RecordBatchIterator;
+use crate::rerankers::RerankHybridCallbackArgs;
 use crate::rerankers::Reranker;
-use crate::rerankers::RerankerCallbacks;
 use crate::util::{parse_distance_type, schema_to_buffer};
 
+fn bytes_to_arrow_array(data: Uint8Array, dtype: String) -> napi::Result<Arc<dyn Array>> {
+    let buf = arrow_buffer::Buffer::from(data.to_vec());
+    let num_bytes = buf.len();
+    match dtype.as_str() {
+        "float16" => {
+            let scalar_buf = ScalarBuffer::<f16>::new(buf, 0, num_bytes / 2);
+            Ok(Arc::new(ArrowFloat16Array::new(scalar_buf, None)))
+        }
+        "float32" => {
+            let scalar_buf = ScalarBuffer::<f32>::new(buf, 0, num_bytes / 4);
+            Ok(Arc::new(ArrowFloat32Array::new(scalar_buf, None)))
+        }
+        "float64" => {
+            let scalar_buf = ScalarBuffer::<f64>::new(buf, 0, num_bytes / 8);
+            Ok(Arc::new(ArrowFloat64Array::new(scalar_buf, None)))
+        }
+        "uint8" => {
+            let scalar_buf = ScalarBuffer::<u8>::new(buf, 0, num_bytes);
+            Ok(Arc::new(ArrowUInt8Array::new(scalar_buf, None)))
+        }
+        _ => Err(napi::Error::from_reason(format!(
+            "Unsupported vector dtype: {}. Expected one of: float16, float32, float64, uint8",
+            dtype
+        ))),
+    }
+}
+
 #[napi]
 pub struct Query {
     inner: LanceDbQuery,
@@ -42,7 +75,7 @@ impl Query {
     }
 
     #[napi]
-    pub fn full_text_search(&mut self, query: napi::JsObject) -> napi::Result<()> {
+    pub fn full_text_search(&mut self, query: Object) -> napi::Result<()> {
         let query = parse_fts_query(query)?;
         self.inner = self.inner.clone().full_text_search(query);
         Ok(())
@@ -78,6 +111,13 @@ impl Query {
         Ok(VectorQuery { inner })
     }
 
+    #[napi]
+    pub fn nearest_to_raw(&mut self, data: Uint8Array, dtype: String) -> Result<VectorQuery> {
+        let array = bytes_to_arrow_array(data, dtype)?;
+        let inner = self.inner.clone().nearest_to(array).default_error()?;
+        Ok(VectorQuery { inner })
+    }
+
     #[napi]
     pub fn fast_search(&mut self) {
         self.inner = self.inner.clone().fast_search();
@@ -163,6 +203,13 @@ impl VectorQuery {
         Ok(())
     }
 
+    #[napi]
+    pub fn add_query_vector_raw(&mut self, data: Uint8Array, dtype: String) -> Result<()> {
+        let array = bytes_to_arrow_array(data, dtype)?;
+        self.inner = self.inner.clone().add_query_vector(array).default_error()?;
+        Ok(())
+    }
+
     #[napi]
     pub fn distance_type(&mut self, distance_type: String) -> napi::Result<()> {
         let distance_type = parse_distance_type(distance_type)?;
@@ -235,7 +282,7 @@ impl VectorQuery {
     }
 
     #[napi]
-    pub fn full_text_search(&mut self, query: napi::JsObject) -> napi::Result<()> {
+    pub fn full_text_search(&mut self, query: Object) -> napi::Result<()> {
         let query = parse_fts_query(query)?;
         self.inner = self.inner.clone().full_text_search(query);
         Ok(())
@@ -272,11 +319,13 @@ impl VectorQuery {
     }
 
     #[napi]
-    pub fn rerank(&mut self, callbacks: RerankerCallbacks) {
+    pub fn rerank(
-        self.inner = self
+        &mut self,
-            .inner
+        rerank_hybrid: Function<RerankHybridCallbackArgs, Promise<Buffer>>,
-            .clone()
+    ) -> napi::Result<()> {
-            .rerank(Arc::new(Reranker::new(callbacks)));
+        let reranker = Reranker::new(rerank_hybrid)?;
+        self.inner = self.inner.clone().rerank(Arc::new(reranker));
+        Ok(())
     }
 
     #[napi(catch_unwind)]
@@ -523,12 +572,12 @@ impl JsFullTextQuery {
     }
 }
 
-fn parse_fts_query(query: napi::JsObject) -> napi::Result<FullTextSearchQuery> {
+fn parse_fts_query(query: Object) -> napi::Result<FullTextSearchQuery> {
-    if let Ok(Some(query)) = query.get::<_, &JsFullTextQuery>("query") {
+    if let Ok(Some(query)) = query.get::<&JsFullTextQuery>("query") {
         Ok(FullTextSearchQuery::new_query(query.inner.clone()))
-    } else if let Ok(Some(query_text)) = query.get::<_, String>("query") {
+    } else if let Ok(Some(query_text)) = query.get::<String>("query") {
         let mut query_text = query_text;
-        let columns = query.get::<_, Option<Vec<String>>>("columns")?.flatten();
+        let columns = query.get::<Option<Vec<String>>>("columns")?.flatten();
 
         let is_phrase =
             query_text.len() >= 2 && query_text.starts_with('"') && query_text.ends_with('"');
@@ -549,15 +598,12 @@ fn parse_fts_query(query: napi::JsObject) -> napi::Result<FullTextSearchQuery> {
             }
         };
         let mut query = FullTextSearchQuery::new_query(query);
-        if let Some(cols) = columns {
+        if let Some(cols) = columns
-            if !cols.is_empty() {
+            && !cols.is_empty()
-                query = query.with_columns(&cols).map_err(|e| {
+        {
-                    napi::Error::from_reason(format!(
+            query = query.with_columns(&cols).map_err(|e| {
-                        "Failed to set full text search columns: {}",
+                napi::Error::from_reason(format!("Failed to set full text search columns: {}", e))
-                        e
+            })?;
-                    ))
-                })?;
-            }
         }
         Ok(query)
     } else {

nodejs/src/rerankers.rs

@@ -3,10 +3,7 @@
 
 use arrow_array::RecordBatch;
 use async_trait::async_trait;
-use napi::{
+use napi::{bindgen_prelude::*, threadsafe_function::ThreadsafeFunction};
-    bindgen_prelude::*,
-    threadsafe_function::{ErrorStrategy, ThreadsafeFunction},
-};
 use napi_derive::napi;
 
 use lancedb::ipc::batches_to_ipc_file;
@@ -15,27 +12,28 @@ use lancedb::{error::Error, ipc::ipc_file_to_batches};
 
 use crate::error::NapiErrorExt;
 
+type RerankHybridFn = ThreadsafeFunction<
+    RerankHybridCallbackArgs,
+    Promise<Buffer>,
+    RerankHybridCallbackArgs,
+    Status,
+    false,
+>;
+
 /// Reranker implementation that "wraps" a NodeJS Reranker implementation.
 /// This contains references to the callbacks that can be used to invoke the
 /// reranking methods on the NodeJS implementation and handles serializing the
 /// record batches to Arrow IPC buffers.
-#[napi]
 pub struct Reranker {
-    /// callback to the Javascript which will call the rerankHybrid method of
+    rerank_hybrid: RerankHybridFn,
-    /// some Reranker implementation
-    rerank_hybrid: ThreadsafeFunction<RerankHybridCallbackArgs, ErrorStrategy::CalleeHandled>,
 }
 
-#[napi]
 impl Reranker {
-    #[napi]
+    pub fn new(
-    pub fn new(callbacks: RerankerCallbacks) -> Self {
+        rerank_hybrid: Function<RerankHybridCallbackArgs, Promise<Buffer>>,
-        let rerank_hybrid = callbacks
+    ) -> napi::Result<Self> {
-            .rerank_hybrid
+        let rerank_hybrid = rerank_hybrid.build_threadsafe_function().build()?;
-            .create_threadsafe_function(0, move |ctx| Ok(vec![ctx.value]))
+        Ok(Self { rerank_hybrid })
-            .unwrap();
-
-        Self { rerank_hybrid }
     }
 }
 
@@ -49,16 +47,16 @@ impl lancedb::rerankers::Reranker for Reranker {
     ) -> lancedb::error::Result<RecordBatch> {
         let callback_args = RerankHybridCallbackArgs {
             query: query.to_string(),
-            vec_results: batches_to_ipc_file(&[vector_results])?,
+            vec_results: Buffer::from(batches_to_ipc_file(&[vector_results])?.as_ref()),
-            fts_results: batches_to_ipc_file(&[fts_results])?,
+            fts_results: Buffer::from(batches_to_ipc_file(&[fts_results])?.as_ref()),
         };
         let promised_buffer: Promise<Buffer> = self
             .rerank_hybrid
-            .call_async(Ok(callback_args))
+            .call_async(callback_args)
             .await
             .map_err(|e| Error::Runtime {
-                message: format!("napi error status={}, reason={}", e.status, e.reason),
+            message: format!("napi error status={}, reason={}", e.status, e.reason),
-            })?;
+        })?;
         let buffer = promised_buffer.await.map_err(|e| Error::Runtime {
             message: format!("napi error status={}, reason={}", e.status, e.reason),
         })?;
@@ -77,16 +75,11 @@ impl std::fmt::Debug for Reranker {
     }
 }
 
-#[napi(object)]
-pub struct RerankerCallbacks {
-    pub rerank_hybrid: JsFunction,
-}
-
 #[napi(object)]
 pub struct RerankHybridCallbackArgs {
     pub query: String,
-    pub vec_results: Vec<u8>,
+    pub vec_results: Buffer,
-    pub fts_results: Vec<u8>,
+    pub fts_results: Buffer,
 }
 
 fn buffer_to_record_batch(buffer: Buffer) -> Result<RecordBatch> {

nodejs/src/session.rs

@@ -95,8 +95,7 @@ impl napi::bindgen_prelude::FromNapiValue for Session {
         napi_val: napi::sys::napi_value,
     ) -> napi::Result<Self> {
         let object: napi::bindgen_prelude::ClassInstance<Self> =
-            napi::bindgen_prelude::ClassInstance::from_napi_value(env, napi_val)?;
+            unsafe { napi::bindgen_prelude::ClassInstance::from_napi_value(env, napi_val)? };
-        let copy = object.clone();
+        Ok((*object).clone())
-        Ok(copy)
     }
 }

nodejs/src/table.rs

@@ -3,7 +3,7 @@
 
 use std::collections::HashMap;
 
-use lancedb::ipc::ipc_file_to_batches;
+use lancedb::ipc::{ipc_file_to_batches, ipc_file_to_schema};
 use lancedb::table::{
     AddDataMode, ColumnAlteration as LanceColumnAlteration, Duration, NewColumnTransform,
     OptimizeAction, OptimizeOptions, Table as LanceDbTable,
@@ -71,6 +71,17 @@ impl Table {
     pub async fn add(&self, buf: Buffer, mode: String) -> napi::Result<AddResult> {
         let batches = ipc_file_to_batches(buf.to_vec())
             .map_err(|e| napi::Error::from_reason(format!("Failed to read IPC file: {}", e)))?;
+        let batches = batches
+            .into_iter()
+            .map(|batch| {
+                batch.map_err(|e| {
+                    napi::Error::from_reason(format!(
+                        "Failed to read record batch from IPC file: {}",
+                        e
+                    ))
+                })
+            })
+            .collect::<Result<Vec<_>>>()?;
         let mut op = self.inner_ref()?.add(batches);
 
         op = if mode == "append" {
@@ -268,6 +279,23 @@ impl Table {
         Ok(res.into())
     }
 
+    #[napi(catch_unwind)]
+    pub async fn add_columns_with_schema(
+        &self,
+        schema_buf: Buffer,
+    ) -> napi::Result<AddColumnsResult> {
+        let schema = ipc_file_to_schema(schema_buf.to_vec())
+            .map_err(|e| napi::Error::from_reason(format!("Failed to read IPC schema: {}", e)))?;
+
+        let transforms = NewColumnTransform::AllNulls(schema);
+        let res = self
+            .inner_ref()?
+            .add_columns(transforms, None)
+            .await
+            .default_error()?;
+        Ok(res.into())
+    }
+
     #[napi(catch_unwind)]
     pub async fn alter_columns(
         &self,
@@ -742,12 +770,14 @@ impl From<lancedb::table::AddResult> for AddResult {
 
 #[napi(object)]
 pub struct DeleteResult {
+    pub num_deleted_rows: i64,
     pub version: i64,
 }
 
 impl From<lancedb::table::DeleteResult> for DeleteResult {
     fn from(value: lancedb::table::DeleteResult) -> Self {
         Self {
+            num_deleted_rows: value.num_deleted_rows as i64,
             version: value.version as i64,
         }
     }

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.30.0-beta.1"
+current_version = "0.31.0-beta.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/.gitignore

@@ -1,3 +1,5 @@
 # Test data created by some example tests
 data/
 _lancedb.pyd
+# macOS debug symbols bundle generated during build
+*.dSYM/

python/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.30.0-beta.1"
+version = "0.31.0-beta.0"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true
@@ -16,11 +16,14 @@ crate-type = ["cdylib"]
 [dependencies]
 arrow = { version = "57.2", features = ["pyarrow"] }
 async-trait = "0.1"
+bytes = "1"
 lancedb = { path = "../rust/lancedb", default-features = false }
 lance-core.workspace = true
 lance-namespace.workspace = true
+lance-namespace-impls.workspace = true
 lance-io.workspace = true
 env_logger.workspace = true
+log.workspace = true
 pyo3 = { version = "0.26", features = ["extension-module", "abi3-py39"] }
 pyo3-async-runtimes = { version = "0.26", features = [
     "attributes",
@@ -28,6 +31,8 @@ pyo3-async-runtimes = { version = "0.26", features = [
 ] }
 pin-project = "1.1.5"
 futures.workspace = true
+serde = "1"
+serde_json = "1"
 snafu.workspace = true
 tokio = { version = "1.40", features = ["sync"] }
 

python/README.md

@@ -1,4 +1,4 @@
-# LanceDB
+# LanceDB Python SDK
 
 A Python library for [LanceDB](https://github.com/lancedb/lancedb).
 

python/pyproject.toml

@@ -3,10 +3,10 @@ name = "lancedb"
 # version in Cargo.toml
 dynamic = ["version"]
 dependencies = [
-    "deprecation",
+    "deprecation>=2.1.0",
-    "numpy",
+    "numpy>=1.24.0",
     "overrides>=0.7; python_version<'3.12'",
-    "packaging",
+    "packaging>=23.0",
     "pyarrow>=16",
     "pydantic>=1.10",
     "tqdm>=4.27.0",
@@ -45,51 +45,51 @@ repository = "https://github.com/lancedb/lancedb"
 
 [project.optional-dependencies]
 pylance = [
-    "pylance>=1.0.0b14",
+    "pylance>=5.0.0b3",
 ]
 tests = [
-    "aiohttp",
+    "aiohttp>=3.9.0",
-    "boto3",
+    "boto3>=1.28.57",
     "pandas>=1.4",
-    "pytest",
+    "pytest>=7.0",
-    "pytest-mock",
+    "pytest-mock>=3.10",
-    "pytest-asyncio",
+    "pytest-asyncio>=0.21",
-    "duckdb",
+    "duckdb>=0.9.0",
-    "pytz",
+    "pytz>=2023.3",
     "polars>=0.19, <=1.3.0",
-    "tantivy",
+    "tantivy>=0.20.0",
-    "pyarrow-stubs",
+    "pyarrow-stubs>=16.0",
-    "pylance>=1.0.0b14",
+    "pylance>=5.0.0b3",
-    "requests",
+    "requests>=2.31.0",
-    "datafusion",
+    "datafusion>=52,<53",
 ]
 dev = [
-    "ruff",
+    "ruff>=0.3.0",
-    "pre-commit",
+    "pre-commit>=3.5.0",
-    "pyright",
+    "pyright>=1.1.350",
     'typing-extensions>=4.0.0; python_version < "3.11"',
 ]
 docs = ["mkdocs", "mkdocs-jupyter", "mkdocs-material", "mkdocstrings-python"]
-clip = ["torch", "pillow", "open-clip-torch"]
+clip = ["torch", "pillow>=12.1.1", "open-clip-torch"]
-siglip = ["torch", "pillow", "transformers>=4.41.0","sentencepiece"]
+siglip = ["torch", "pillow>=12.1.1", "transformers>=4.41.0","sentencepiece"]
 embeddings = [
     "requests>=2.31.0",
     "openai>=1.6.1",
-    "sentence-transformers",
+    "sentence-transformers>=2.2.0",
-    "torch",
+    "torch>=2.0.0",
-    "pillow",
+    "pillow>=12.1.1",
-    "open-clip-torch",
+    "open-clip-torch>=2.20.0",
-    "cohere",
+    "cohere>=4.0",
     "colpali-engine>=0.3.10",
-    "huggingface_hub",
+    "huggingface_hub>=0.19.0",
-    "InstructorEmbedding",
+    "InstructorEmbedding>=1.0.1",
-    "google.generativeai",
+    "google.generativeai>=0.3.0",
     "boto3>=1.28.57",
-    "awscli>=1.29.57",
+    "awscli>=1.44.38",
     "botocore>=1.31.57",
     'ibm-watsonx-ai>=1.1.2; python_version >= "3.10"',
     "ollama>=0.3.0",
-    "sentencepiece"
+    "sentencepiece>=0.1.99"
 ]
 azure = ["adlfs>=2024.2.0"]
 

python/python/lancedb/__init__.py

@@ -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
+from typing import Dict, Optional, Union, Any, List
 import warnings
 
 __version__ = importlib.metadata.version("lancedb")
@@ -15,9 +15,9 @@ from ._lancedb import connect as lancedb_connect
 from .common import URI, sanitize_uri
 from urllib.parse import urlparse
 from .db import AsyncConnection, DBConnection, LanceDBConnection
-from .io import StorageOptionsProvider
 from .remote import ClientConfig
 from .remote.db import RemoteDBConnection
+from .expr import Expr, col, lit, func
 from .schema import vector
 from .table import AsyncTable, Table
 from ._lancedb import Session
@@ -63,7 +63,7 @@ def _check_s3_bucket_with_dots(
 
 
 def connect(
-    uri: URI,
+    uri: Optional[URI] = None,
     *,
     api_key: Optional[str] = None,
     region: str = "us-east-1",
@@ -73,14 +73,18 @@ def connect(
     client_config: Union[ClientConfig, Dict[str, Any], None] = None,
     storage_options: Optional[Dict[str, str]] = None,
     session: Optional[Session] = None,
+    namespace_client_impl: Optional[str] = None,
+    namespace_client_properties: Optional[Dict[str, str]] = None,
+    namespace_client_pushdown_operations: Optional[List[str]] = None,
     **kwargs: Any,
 ) -> DBConnection:
     """Connect to a LanceDB database.
 
     Parameters
     ----------
-    uri: str or Path
+    uri: str or Path, optional
-        The uri of the database.
+        The uri of the database. When ``namespace_client_impl`` is provided you may
+        omit ``uri`` and connect through a namespace client instead.
     api_key: str, optional
         If presented, connect to LanceDB cloud.
         Otherwise, connect to a database on file system or cloud storage.
@@ -113,6 +117,18 @@ def connect(
         cache sizes for index and metadata caches, which can significantly
         impact memory use and performance. They can also be re-used across
         multiple connections to share the same cache state.
+    namespace_client_impl : str, optional
+        When provided along with ``namespace_client_properties``, ``connect``
+        returns a namespace-backed connection by delegating to
+        :func:`connect_namespace`. The value identifies which namespace
+        implementation to load (e.g., ``"dir"`` or ``"rest"``).
+    namespace_client_properties : dict, optional
+        Configuration to pass to the namespace client implementation. Required
+        when ``namespace_client_impl`` is set.
+    namespace_client_pushdown_operations : list[str], optional
+        Only used when ``namespace_client_properties`` is provided. Forwards to
+        :func:`connect_namespace` to control which operations are executed on the
+        namespace service (e.g., ``["QueryTable", "CreateTable"]``).
 
     Examples
     --------
@@ -132,11 +148,42 @@ def connect(
     >>> db = lancedb.connect("db://my_database", api_key="ldb_...",
     ...                      client_config={"retry_config": {"retries": 5}})
 
+    Connect to a namespace-backed database:
+
+    >>> db = lancedb.connect(namespace_client_impl="dir",
+    ...                      namespace_client_properties={"root": "/tmp/ns"})
+
     Returns
     -------
     conn : DBConnection
         A connection to a LanceDB database.
     """
+    if namespace_client_impl is not None or namespace_client_properties is not None:
+        if namespace_client_impl is None or namespace_client_properties is None:
+            raise ValueError(
+                "Both namespace_client_impl and "
+                "namespace_client_properties must be provided"
+            )
+        if kwargs:
+            raise ValueError(f"Unknown keyword arguments: {kwargs}")
+        return connect_namespace(
+            namespace_client_impl,
+            namespace_client_properties,
+            read_consistency_interval=read_consistency_interval,
+            storage_options=storage_options,
+            session=session,
+            namespace_client_pushdown_operations=namespace_client_pushdown_operations,
+        )
+
+    if namespace_client_pushdown_operations is not None:
+        raise ValueError(
+            "namespace_client_pushdown_operations is only valid when "
+            "connecting through a namespace"
+        )
+    if uri is None:
+        raise ValueError(
+            "uri is required when not connecting through a namespace client"
+        )
     if isinstance(uri, str) and uri.startswith("db://"):
         if api_key is None:
             api_key = os.environ.get("LANCEDB_API_KEY")
@@ -271,6 +318,10 @@ __all__ = [
     "AsyncConnection",
     "AsyncLanceNamespaceDBConnection",
     "AsyncTable",
+    "col",
+    "Expr",
+    "func",
+    "lit",
     "URI",
     "sanitize_uri",
     "vector",
@@ -279,7 +330,6 @@ __all__ = [
     "LanceNamespaceDBConnection",
     "RemoteDBConnection",
     "Session",
-    "StorageOptionsProvider",
     "Table",
     "__version__",
 ]

python/python/lancedb/_lancedb.pyi

@@ -14,7 +14,6 @@ from .index import (
     HnswSq,
     FTS,
 )
-from .io import StorageOptionsProvider
 from lance_namespace import (
     ListNamespacesResponse,
     CreateNamespaceResponse,
@@ -27,6 +26,32 @@ from .remote import ClientConfig
 IvfHnswPq: type[HnswPq] = HnswPq
 IvfHnswSq: type[HnswSq] = HnswSq
 
+class PyExpr:
+    """A type-safe DataFusion expression node (Rust-side handle)."""
+
+    def eq(self, other: "PyExpr") -> "PyExpr": ...
+    def ne(self, other: "PyExpr") -> "PyExpr": ...
+    def lt(self, other: "PyExpr") -> "PyExpr": ...
+    def lte(self, other: "PyExpr") -> "PyExpr": ...
+    def gt(self, other: "PyExpr") -> "PyExpr": ...
+    def gte(self, other: "PyExpr") -> "PyExpr": ...
+    def and_(self, other: "PyExpr") -> "PyExpr": ...
+    def or_(self, other: "PyExpr") -> "PyExpr": ...
+    def not_(self) -> "PyExpr": ...
+    def add(self, other: "PyExpr") -> "PyExpr": ...
+    def sub(self, other: "PyExpr") -> "PyExpr": ...
+    def mul(self, other: "PyExpr") -> "PyExpr": ...
+    def div(self, other: "PyExpr") -> "PyExpr": ...
+    def lower(self) -> "PyExpr": ...
+    def upper(self) -> "PyExpr": ...
+    def contains(self, substr: "PyExpr") -> "PyExpr": ...
+    def cast(self, data_type: pa.DataType) -> "PyExpr": ...
+    def to_sql(self) -> str: ...
+
+def expr_col(name: str) -> PyExpr: ...
+def expr_lit(value: Union[bool, int, float, str]) -> PyExpr: ...
+def expr_func(name: str, args: List[PyExpr]) -> PyExpr: ...
+
 class Session:
     def __init__(
         self,
@@ -46,35 +71,35 @@ class Connection(object):
     async def close(self): ...
     async def list_namespaces(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListNamespacesResponse: ...
     async def create_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         properties: Optional[Dict[str, str]] = None,
     ) -> CreateNamespaceResponse: ...
     async def drop_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         behavior: Optional[str] = None,
     ) -> DropNamespaceResponse: ...
     async def describe_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
     ) -> DescribeNamespaceResponse: ...
     async def list_tables(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListTablesResponse: ...
     async def table_names(
         self,
-        namespace: Optional[List[str]],
+        namespace_path: Optional[List[str]],
         start_after: Optional[str],
         limit: Optional[int],
     ) -> list[str]: ...  # Deprecated: Use list_tables instead
@@ -83,9 +108,8 @@ class Connection(object):
         name: str,
         mode: str,
         data: pa.RecordBatchReader,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional[StorageOptionsProvider] = None,
         location: Optional[str] = None,
     ) -> Table: ...
     async def create_empty_table(
@@ -93,17 +117,15 @@ class Connection(object):
         name: str,
         mode: str,
         schema: pa.Schema,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional[StorageOptionsProvider] = None,
         location: Optional[str] = None,
     ) -> Table: ...
     async def open_table(
         self,
         name: str,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional[StorageOptionsProvider] = None,
         index_cache_size: Optional[int] = None,
         location: Optional[str] = None,
     ) -> Table: ...
@@ -111,7 +133,7 @@ class Connection(object):
         self,
         target_table_name: str,
         source_uri: str,
-        target_namespace: Optional[List[str]] = None,
+        target_namespace_path: Optional[List[str]] = None,
         source_version: Optional[int] = None,
         source_tag: Optional[str] = None,
         is_shallow: bool = True,
@@ -120,13 +142,15 @@ class Connection(object):
         self,
         cur_name: str,
         new_name: str,
-        cur_namespace: Optional[List[str]] = None,
+        cur_namespace_path: Optional[List[str]] = None,
-        new_namespace: Optional[List[str]] = None,
+        new_namespace_path: Optional[List[str]] = None,
     ) -> None: ...
     async def drop_table(
-        self, name: str, namespace: Optional[List[str]] = None
+        self, name: str, namespace_path: Optional[List[str]] = None
+    ) -> None: ...
+    async def drop_all_tables(
+        self, namespace_path: Optional[List[str]] = None
     ) -> None: ...
-    async def drop_all_tables(self, namespace: Optional[List[str]] = None) -> None: ...
 
 class Table:
     def name(self) -> str: ...
@@ -135,7 +159,10 @@ class Table:
     def close(self) -> None: ...
     async def schema(self) -> pa.Schema: ...
     async def add(
-        self, data: pa.RecordBatchReader, mode: Literal["append", "overwrite"]
+        self,
+        data: pa.RecordBatchReader,
+        mode: Literal["append", "overwrite"],
+        progress: Optional[Any] = None,
     ) -> AddResult: ...
     async def update(
         self, updates: Dict[str, str], where: Optional[str]
@@ -166,6 +193,8 @@ class Table:
     async def checkout(self, version: Union[int, str]): ...
     async def checkout_latest(self): ...
     async def restore(self, version: Optional[Union[int, str]] = None): ...
+    async def prewarm_index(self, index_name: str) -> None: ...
+    async def prewarm_data(self, columns: Optional[List[str]] = None) -> None: ...
     async def list_indices(self) -> list[IndexConfig]: ...
     async def delete(self, filter: str) -> DeleteResult: ...
     async def add_columns(self, columns: list[tuple[str, str]]) -> AddColumnsResult: ...
@@ -220,7 +249,9 @@ class RecordBatchStream:
 
 class Query:
     def where(self, filter: str): ...
-    def select(self, columns: Tuple[str, str]): ...
+    def where_expr(self, expr: PyExpr): ...
+    def select(self, columns: List[Tuple[str, str]]): ...
+    def select_expr(self, columns: List[Tuple[str, PyExpr]]): ...
     def select_columns(self, columns: List[str]): ...
     def limit(self, limit: int): ...
     def offset(self, offset: int): ...
@@ -246,7 +277,9 @@ class TakeQuery:
 
 class FTSQuery:
     def where(self, filter: str): ...
-    def select(self, columns: List[str]): ...
+    def where_expr(self, expr: PyExpr): ...
+    def select(self, columns: List[Tuple[str, str]]): ...
+    def select_expr(self, columns: List[Tuple[str, PyExpr]]): ...
     def limit(self, limit: int): ...
     def offset(self, offset: int): ...
     def fast_search(self): ...
@@ -265,7 +298,9 @@ class VectorQuery:
     async def output_schema(self) -> pa.Schema: ...
     async def execute(self) -> RecordBatchStream: ...
     def where(self, filter: str): ...
-    def select(self, columns: List[str]): ...
+    def where_expr(self, expr: PyExpr): ...
+    def select(self, columns: List[Tuple[str, str]]): ...
+    def select_expr(self, columns: List[Tuple[str, PyExpr]]): ...
     def select_with_projection(self, columns: Tuple[str, str]): ...
     def limit(self, limit: int): ...
     def offset(self, offset: int): ...
@@ -282,7 +317,9 @@ class VectorQuery:
 
 class HybridQuery:
     def where(self, filter: str): ...
-    def select(self, columns: List[str]): ...
+    def where_expr(self, expr: PyExpr): ...
+    def select(self, columns: List[Tuple[str, str]]): ...
+    def select_expr(self, columns: List[Tuple[str, PyExpr]]): ...
     def limit(self, limit: int): ...
     def offset(self, offset: int): ...
     def fast_search(self): ...

python/python/lancedb/arrow.py

@@ -1,8 +1,10 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
+from functools import singledispatch
 from typing import List, Optional, Tuple, Union
 
+from lancedb.pydantic import LanceModel, model_to_dict
 import pyarrow as pa
 
 from ._lancedb import RecordBatchStream
@@ -80,3 +82,32 @@ def peek_reader(
         yield from reader
 
     return batch, pa.RecordBatchReader.from_batches(batch.schema, all_batches())
+
+
+@singledispatch
+def to_arrow(data) -> pa.Table:
+    """Convert a single data object to a pa.Table."""
+    raise NotImplementedError(f"to_arrow not implemented for type {type(data)}")
+
+
+@to_arrow.register(pa.RecordBatch)
+def _arrow_from_batch(data: pa.RecordBatch) -> pa.Table:
+    return pa.Table.from_batches([data])
+
+
+@to_arrow.register(pa.Table)
+def _arrow_from_table(data: pa.Table) -> pa.Table:
+    return data
+
+
+@to_arrow.register(list)
+def _arrow_from_list(data: list) -> pa.Table:
+    if not data:
+        raise ValueError("Cannot create table from empty list without a schema")
+
+    if isinstance(data[0], LanceModel):
+        schema = data[0].__class__.to_arrow_schema()
+        dicts = [model_to_dict(d) for d in data]
+        return pa.Table.from_pylist(dicts, schema=schema)
+
+    return pa.Table.from_pylist(data)

python/python/lancedb/db.py

@@ -8,7 +8,7 @@ from abc import abstractmethod
 from datetime import timedelta
 from pathlib import Path
 import sys
-from typing import TYPE_CHECKING, Dict, Iterable, List, Literal, Optional, Union
+from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Literal, Optional, Union
 
 if sys.version_info >= (3, 12):
     from typing import override
@@ -52,7 +52,6 @@ if TYPE_CHECKING:
     from ._lancedb import Connection as LanceDbConnection
     from .common import DATA, URI
     from .embeddings import EmbeddingFunctionConfig
-    from .io import StorageOptionsProvider
     from ._lancedb import Session
 
 from .namespace_utils import (
@@ -67,7 +66,7 @@ class DBConnection(EnforceOverrides):
 
     def list_namespaces(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListNamespacesResponse:
@@ -75,7 +74,7 @@ class DBConnection(EnforceOverrides):
 
         Parameters
         ----------
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The parent namespace to list namespaces in.
             Empty list represents root namespace.
         page_token: str, optional
@@ -89,13 +88,13 @@ class DBConnection(EnforceOverrides):
         ListNamespacesResponse
             Response containing namespace names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return ListNamespacesResponse(namespaces=[], page_token=None)
 
     def create_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         properties: Optional[Dict[str, str]] = None,
     ) -> CreateNamespaceResponse:
@@ -103,7 +102,7 @@ class DBConnection(EnforceOverrides):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to create.
         mode: str, optional
             Creation mode - "create" (fail if exists), "exist_ok" (skip if exists),
@@ -122,7 +121,7 @@ class DBConnection(EnforceOverrides):
 
     def drop_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         behavior: Optional[str] = None,
     ) -> DropNamespaceResponse:
@@ -130,7 +129,7 @@ class DBConnection(EnforceOverrides):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to drop.
         mode: str, optional
             Whether to skip if not exists ("SKIP") or fail ("FAIL"). Case insensitive.
@@ -147,12 +146,14 @@ class DBConnection(EnforceOverrides):
             "Namespace operations are not supported for this connection type"
         )
 
-    def describe_namespace(self, namespace: List[str]) -> DescribeNamespaceResponse:
+    def describe_namespace(
+        self, namespace_path: List[str]
+    ) -> DescribeNamespaceResponse:
         """Describe a namespace.
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to describe.
 
         Returns
@@ -166,7 +167,7 @@ class DBConnection(EnforceOverrides):
 
     def list_tables(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListTablesResponse:
@@ -174,7 +175,7 @@ class DBConnection(EnforceOverrides):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -198,13 +199,13 @@ class DBConnection(EnforceOverrides):
         page_token: Optional[str] = None,
         limit: int = 10,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
     ) -> Iterable[str]:
         """List all tables in this database, in sorted order
 
         Parameters
         ----------
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to list tables in.
             Empty list represents root namespace.
         page_token: str, optional
@@ -231,9 +232,8 @@ class DBConnection(EnforceOverrides):
         fill_value: float = 0.0,
         embedding_functions: Optional[List[EmbeddingFunctionConfig]] = None,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         data_storage_version: Optional[str] = None,
         enable_v2_manifest_paths: Optional[bool] = None,
     ) -> Table:
@@ -243,7 +243,7 @@ class DBConnection(EnforceOverrides):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to create the table in.
             Empty list represents root namespace.
         data: The data to initialize the table, *optional*
@@ -401,9 +401,8 @@ class DBConnection(EnforceOverrides):
         self,
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         index_cache_size: Optional[int] = None,
     ) -> Table:
         """Open a Lance Table in the database.
@@ -412,7 +411,7 @@ class DBConnection(EnforceOverrides):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to open the table from.
             None or empty list represents root namespace.
         index_cache_size: int, default 256
@@ -440,27 +439,27 @@ class DBConnection(EnforceOverrides):
         """
         raise NotImplementedError
 
-    def drop_table(self, name: str, namespace: Optional[List[str]] = None):
+    def drop_table(self, name: str, namespace_path: Optional[List[str]] = None):
         """Drop a table from the database.
 
         Parameters
         ----------
         name: str
             The name of the table.
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to drop the table from.
             Empty list represents root namespace.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         raise NotImplementedError
 
     def rename_table(
         self,
         cur_name: str,
         new_name: str,
-        cur_namespace: Optional[List[str]] = None,
+        cur_namespace_path: Optional[List[str]] = None,
-        new_namespace: Optional[List[str]] = None,
+        new_namespace_path: Optional[List[str]] = None,
     ):
         """Rename a table in the database.
 
@@ -470,17 +469,17 @@ class DBConnection(EnforceOverrides):
             The current name of the table.
         new_name: str
             The new name of the table.
-        cur_namespace: List[str], optional
+        cur_namespace_path: List[str], optional
             The namespace of the current table.
             None or empty list represents root namespace.
-        new_namespace: List[str], optional
+        new_namespace_path: List[str], optional
             The namespace to move the table to.
             If not specified, defaults to the same as cur_namespace.
         """
-        if cur_namespace is None:
+        if cur_namespace_path is None:
-            cur_namespace = []
+            cur_namespace_path = []
-        if new_namespace is None:
+        if new_namespace_path is None:
-            new_namespace = []
+            new_namespace_path = []
         raise NotImplementedError
 
     def drop_database(self):
@@ -490,18 +489,18 @@ class DBConnection(EnforceOverrides):
         """
         raise NotImplementedError
 
-    def drop_all_tables(self, namespace: Optional[List[str]] = None):
+    def drop_all_tables(self, namespace_path: Optional[List[str]] = None):
         """
         Drop all tables from the database
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to drop all tables from.
             None or empty list represents root namespace.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         raise NotImplementedError
 
     @property
@@ -642,7 +641,7 @@ class LanceDBConnection(DBConnection):
     @override
     def list_namespaces(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListNamespacesResponse:
@@ -650,7 +649,7 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The parent namespace to list namespaces in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -664,18 +663,18 @@ class LanceDBConnection(DBConnection):
         ListNamespacesResponse
             Response containing namespace names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.list_namespaces(
-                namespace=namespace, page_token=page_token, limit=limit
+                namespace_path=namespace_path, page_token=page_token, limit=limit
             )
         )
 
     @override
     def create_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         properties: Optional[Dict[str, str]] = None,
     ) -> CreateNamespaceResponse:
@@ -683,7 +682,7 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to create.
         mode: str, optional
             Creation mode - "create" (fail if exists), "exist_ok" (skip if exists),
@@ -698,14 +697,14 @@ class LanceDBConnection(DBConnection):
         """
         return LOOP.run(
             self._conn.create_namespace(
-                namespace=namespace, mode=mode, properties=properties
+                namespace_path=namespace_path, mode=mode, properties=properties
             )
         )
 
     @override
     def drop_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         behavior: Optional[str] = None,
     ) -> DropNamespaceResponse:
@@ -713,7 +712,7 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to drop.
         mode: str, optional
             Whether to skip if not exists ("SKIP") or fail ("FAIL"). Case insensitive.
@@ -727,16 +726,20 @@ class LanceDBConnection(DBConnection):
             Response containing properties and transaction_id if applicable.
         """
         return LOOP.run(
-            self._conn.drop_namespace(namespace=namespace, mode=mode, behavior=behavior)
+            self._conn.drop_namespace(
+                namespace_path=namespace_path, mode=mode, behavior=behavior
+            )
         )
 
     @override
-    def describe_namespace(self, namespace: List[str]) -> DescribeNamespaceResponse:
+    def describe_namespace(
+        self, namespace_path: List[str]
+    ) -> DescribeNamespaceResponse:
         """Describe a namespace.
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to describe.
 
         Returns
@@ -744,12 +747,12 @@ class LanceDBConnection(DBConnection):
         DescribeNamespaceResponse
             Response containing the namespace properties.
         """
-        return LOOP.run(self._conn.describe_namespace(namespace=namespace))
+        return LOOP.run(self._conn.describe_namespace(namespace_path=namespace_path))
 
     @override
     def list_tables(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListTablesResponse:
@@ -757,7 +760,7 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -771,11 +774,11 @@ class LanceDBConnection(DBConnection):
         ListTablesResponse
             Response containing table names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.list_tables(
-                namespace=namespace, page_token=page_token, limit=limit
+                namespace_path=namespace_path, page_token=page_token, limit=limit
             )
         )
 
@@ -785,7 +788,7 @@ class LanceDBConnection(DBConnection):
         page_token: Optional[str] = None,
         limit: int = 10,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
     ) -> Iterable[str]:
         """Get the names of all tables in the database. The names are sorted.
 
@@ -794,7 +797,7 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
         page_token: str, optional
             The token to use for pagination.
@@ -813,11 +816,11 @@ class LanceDBConnection(DBConnection):
             DeprecationWarning,
             stacklevel=2,
         )
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.table_names(
-                namespace=namespace, start_after=page_token, limit=limit
+                namespace_path=namespace_path, start_after=page_token, limit=limit
             )
         )
 
@@ -839,9 +842,8 @@ class LanceDBConnection(DBConnection):
         fill_value: float = 0.0,
         embedding_functions: Optional[List[EmbeddingFunctionConfig]] = None,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         data_storage_version: Optional[str] = None,
         enable_v2_manifest_paths: Optional[bool] = None,
     ) -> LanceTable:
@@ -849,15 +851,15 @@ class LanceDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to create the table in.
 
         See
         ---
         DBConnection.create_table
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         if mode.lower() not in ["create", "overwrite"]:
             raise ValueError("mode must be either 'create' or 'overwrite'")
         validate_table_name(name)
@@ -872,9 +874,8 @@ class LanceDBConnection(DBConnection):
             on_bad_vectors=on_bad_vectors,
             fill_value=fill_value,
             embedding_functions=embedding_functions,
-            namespace=namespace,
+            namespace_path=namespace_path,
             storage_options=storage_options,
-            storage_options_provider=storage_options_provider,
         )
         return tbl
 
@@ -883,9 +884,8 @@ class LanceDBConnection(DBConnection):
         self,
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         index_cache_size: Optional[int] = None,
     ) -> LanceTable:
         """Open a table in the database.
@@ -894,15 +894,15 @@ class LanceDBConnection(DBConnection):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to open the table from.
 
         Returns
         -------
         A LanceTable object representing the table.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         if index_cache_size is not None:
             import warnings
 
@@ -917,9 +917,8 @@ class LanceDBConnection(DBConnection):
         return LanceTable.open(
             self,
             name,
-            namespace=namespace,
+            namespace_path=namespace_path,
             storage_options=storage_options,
-            storage_options_provider=storage_options_provider,
             index_cache_size=index_cache_size,
         )
 
@@ -928,7 +927,7 @@ class LanceDBConnection(DBConnection):
         target_table_name: str,
         source_uri: str,
         *,
-        target_namespace: Optional[List[str]] = None,
+        target_namespace_path: Optional[List[str]] = None,
         source_version: Optional[int] = None,
         source_tag: Optional[str] = None,
         is_shallow: bool = True,
@@ -946,7 +945,7 @@ class LanceDBConnection(DBConnection):
             The name of the target table to create.
         source_uri: str
             The URI of the source table to clone from.
-        target_namespace: List[str], optional
+        target_namespace_path: List[str], optional
             The namespace for the target table.
             None or empty list represents root namespace.
         source_version: int, optional
@@ -961,13 +960,13 @@ class LanceDBConnection(DBConnection):
         -------
         A LanceTable object representing the cloned table.
         """
-        if target_namespace is None:
+        if target_namespace_path is None:
-            target_namespace = []
+            target_namespace_path = []
         LOOP.run(
             self._conn.clone_table(
                 target_table_name,
                 source_uri,
-                target_namespace=target_namespace,
+                target_namespace_path=target_namespace_path,
                 source_version=source_version,
                 source_tag=source_tag,
                 is_shallow=is_shallow,
@@ -976,14 +975,14 @@ class LanceDBConnection(DBConnection):
         return LanceTable.open(
             self,
             target_table_name,
-            namespace=target_namespace,
+            namespace_path=target_namespace_path,
         )
 
     @override
     def drop_table(
         self,
         name: str,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         ignore_missing: bool = False,
     ):
         """Drop a table from the database.
@@ -992,32 +991,32 @@ class LanceDBConnection(DBConnection):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to drop the table from.
         ignore_missing: bool, default False
             If True, ignore if the table does not exist.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         LOOP.run(
             self._conn.drop_table(
-                name, namespace=namespace, ignore_missing=ignore_missing
+                name, namespace_path=namespace_path, ignore_missing=ignore_missing
             )
         )
 
     @override
-    def drop_all_tables(self, namespace: Optional[List[str]] = None):
+    def drop_all_tables(self, namespace_path: Optional[List[str]] = None):
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
-        LOOP.run(self._conn.drop_all_tables(namespace=namespace))
+        LOOP.run(self._conn.drop_all_tables(namespace_path=namespace_path))
 
     @override
     def rename_table(
         self,
         cur_name: str,
         new_name: str,
-        cur_namespace: Optional[List[str]] = None,
+        cur_namespace_path: Optional[List[str]] = None,
-        new_namespace: Optional[List[str]] = None,
+        new_namespace_path: Optional[List[str]] = None,
     ):
         """Rename a table in the database.
 
@@ -1027,21 +1026,21 @@ class LanceDBConnection(DBConnection):
             The current name of the table.
         new_name: str
             The new name of the table.
-        cur_namespace: List[str], optional
+        cur_namespace_path: List[str], optional
             The namespace of the current table.
-        new_namespace: List[str], optional
+        new_namespace_path: List[str], optional
             The namespace to move the table to.
         """
-        if cur_namespace is None:
+        if cur_namespace_path is None:
-            cur_namespace = []
+            cur_namespace_path = []
-        if new_namespace is None:
+        if new_namespace_path is None:
-            new_namespace = []
+            new_namespace_path = []
         LOOP.run(
             self._conn.rename_table(
                 cur_name,
                 new_name,
-                cur_namespace=cur_namespace,
+                cur_namespace_path=cur_namespace_path,
-                new_namespace=new_namespace,
+                new_namespace_path=new_namespace_path,
             )
         )
 
@@ -1125,7 +1124,7 @@ class AsyncConnection(object):
 
     async def list_namespaces(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListNamespacesResponse:
@@ -1133,7 +1132,7 @@ class AsyncConnection(object):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The parent namespace to list namespaces in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -1146,16 +1145,16 @@ class AsyncConnection(object):
         ListNamespacesResponse
             Response containing namespace names and optional pagination token
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         result = await self._inner.list_namespaces(
-            namespace=namespace, page_token=page_token, limit=limit
+            namespace_path=namespace_path, page_token=page_token, limit=limit
         )
         return ListNamespacesResponse(**result)
 
     async def create_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         properties: Optional[Dict[str, str]] = None,
     ) -> CreateNamespaceResponse:
@@ -1163,7 +1162,7 @@ class AsyncConnection(object):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to create.
         mode: str, optional
             Creation mode - "create", "exist_ok", or "overwrite". Case insensitive.
@@ -1176,7 +1175,7 @@ class AsyncConnection(object):
             Response containing namespace properties
         """
         result = await self._inner.create_namespace(
-            namespace,
+            namespace_path,
             mode=_normalize_create_namespace_mode(mode),
             properties=properties,
         )
@@ -1184,7 +1183,7 @@ class AsyncConnection(object):
 
     async def drop_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         behavior: Optional[str] = None,
     ) -> DropNamespaceResponse:
@@ -1192,7 +1191,7 @@ class AsyncConnection(object):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to drop.
         mode: str, optional
             Whether to skip if not exists ("SKIP") or fail ("FAIL"). Case insensitive.
@@ -1206,20 +1205,20 @@ class AsyncConnection(object):
             Response containing properties and transaction_id if applicable.
         """
         result = await self._inner.drop_namespace(
-            namespace,
+            namespace_path,
             mode=_normalize_drop_namespace_mode(mode),
             behavior=_normalize_drop_namespace_behavior(behavior),
         )
         return DropNamespaceResponse(**result)
 
     async def describe_namespace(
-        self, namespace: List[str]
+        self, namespace_path: List[str]
     ) -> DescribeNamespaceResponse:
         """Describe a namespace.
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to describe.
 
         Returns
@@ -1227,12 +1226,12 @@ class AsyncConnection(object):
         DescribeNamespaceResponse
             Response containing the namespace properties.
         """
-        result = await self._inner.describe_namespace(namespace)
+        result = await self._inner.describe_namespace(namespace_path)
         return DescribeNamespaceResponse(**result)
 
     async def list_tables(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListTablesResponse:
@@ -1240,7 +1239,7 @@ class AsyncConnection(object):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -1254,17 +1253,17 @@ class AsyncConnection(object):
         ListTablesResponse
             Response containing table names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         result = await self._inner.list_tables(
-            namespace=namespace, page_token=page_token, limit=limit
+            namespace_path=namespace_path, page_token=page_token, limit=limit
         )
         return ListTablesResponse(**result)
 
     async def table_names(
         self,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         start_after: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> Iterable[str]:
@@ -1275,7 +1274,7 @@ class AsyncConnection(object):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
             None or empty list represents root namespace.
         start_after: str, optional
@@ -1298,10 +1297,10 @@ class AsyncConnection(object):
             DeprecationWarning,
             stacklevel=2,
         )
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return await self._inner.table_names(
-            namespace=namespace, start_after=start_after, limit=limit
+            namespace_path=namespace_path, start_after=start_after, limit=limit
         )
 
     async def create_table(
@@ -1314,9 +1313,8 @@ class AsyncConnection(object):
         on_bad_vectors: Optional[str] = None,
         fill_value: Optional[float] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         embedding_functions: Optional[List[EmbeddingFunctionConfig]] = None,
         location: Optional[str] = None,
     ) -> AsyncTable:
@@ -1326,7 +1324,7 @@ class AsyncConnection(object):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to create the table in.
             Empty list represents root namespace.
         data: The data to initialize the table, *optional*
@@ -1477,8 +1475,8 @@ class AsyncConnection(object):
         ...     await db.create_table("table4", make_batches(), schema=schema)
         >>> asyncio.run(iterable_example())
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         metadata = None
 
         if embedding_functions is not None:
@@ -1513,9 +1511,8 @@ class AsyncConnection(object):
                 name,
                 mode,
                 schema,
-                namespace=namespace,
+                namespace_path=namespace_path,
                 storage_options=storage_options,
-                storage_options_provider=storage_options_provider,
                 location=location,
             )
         else:
@@ -1524,9 +1521,8 @@ class AsyncConnection(object):
                 name,
                 mode,
                 data,
-                namespace=namespace,
+                namespace_path=namespace_path,
                 storage_options=storage_options,
-                storage_options_provider=storage_options_provider,
                 location=location,
             )
 
@@ -1536,11 +1532,12 @@ class AsyncConnection(object):
         self,
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         index_cache_size: Optional[int] = None,
         location: Optional[str] = None,
+        namespace_client: Optional[Any] = None,
+        managed_versioning: Optional[bool] = None,
     ) -> AsyncTable:
         """Open a Lance Table in the database.
 
@@ -1548,7 +1545,7 @@ class AsyncConnection(object):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to open the table from.
             None or empty list represents root namespace.
         storage_options: dict, optional
@@ -1573,20 +1570,24 @@ class AsyncConnection(object):
             The explicit location (URI) of the table. If provided, the table will be
             opened from this location instead of deriving it from the database URI
             and table name.
+        managed_versioning: bool, optional
+            Whether managed versioning is enabled for this table. If provided,
+            avoids a redundant describe_table call when namespace_client is set.
 
         Returns
         -------
         A LanceTable object representing the table.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         table = await self._inner.open_table(
             name,
-            namespace=namespace,
+            namespace_path=namespace_path,
             storage_options=storage_options,
-            storage_options_provider=storage_options_provider,
             index_cache_size=index_cache_size,
             location=location,
+            namespace_client=namespace_client,
+            managed_versioning=managed_versioning,
         )
         return AsyncTable(table)
 
@@ -1595,7 +1596,7 @@ class AsyncConnection(object):
         target_table_name: str,
         source_uri: str,
         *,
-        target_namespace: Optional[List[str]] = None,
+        target_namespace_path: Optional[List[str]] = None,
         source_version: Optional[int] = None,
         source_tag: Optional[str] = None,
         is_shallow: bool = True,
@@ -1613,7 +1614,7 @@ class AsyncConnection(object):
             The name of the target table to create.
         source_uri: str
             The URI of the source table to clone from.
-        target_namespace: List[str], optional
+        target_namespace_path: List[str], optional
             The namespace for the target table.
             None or empty list represents root namespace.
         source_version: int, optional
@@ -1628,12 +1629,12 @@ class AsyncConnection(object):
         -------
         An AsyncTable object representing the cloned table.
         """
-        if target_namespace is None:
+        if target_namespace_path is None:
-            target_namespace = []
+            target_namespace_path = []
         table = await self._inner.clone_table(
             target_table_name,
             source_uri,
-            target_namespace=target_namespace,
+            target_namespace_path=target_namespace_path,
             source_version=source_version,
             source_tag=source_tag,
             is_shallow=is_shallow,
@@ -1644,8 +1645,8 @@ class AsyncConnection(object):
         self,
         cur_name: str,
         new_name: str,
-        cur_namespace: Optional[List[str]] = None,
+        cur_namespace_path: Optional[List[str]] = None,
-        new_namespace: Optional[List[str]] = None,
+        new_namespace_path: Optional[List[str]] = None,
     ):
         """Rename a table in the database.
 
@@ -1655,26 +1656,29 @@ class AsyncConnection(object):
             The current name of the table.
         new_name: str
             The new name of the table.
-        cur_namespace: List[str], optional
+        cur_namespace_path: List[str], optional
             The namespace of the current table.
             None or empty list represents root namespace.
-        new_namespace: List[str], optional
+        new_namespace_path: List[str], optional
             The namespace to move the table to.
             If not specified, defaults to the same as cur_namespace.
         """
-        if cur_namespace is None:
+        if cur_namespace_path is None:
-            cur_namespace = []
+            cur_namespace_path = []
-        if new_namespace is None:
+        if new_namespace_path is None:
-            new_namespace = []
+            new_namespace_path = []
         await self._inner.rename_table(
-            cur_name, new_name, cur_namespace=cur_namespace, new_namespace=new_namespace
+            cur_name,
+            new_name,
+            cur_namespace_path=cur_namespace_path,
+            new_namespace_path=new_namespace_path,
         )
 
     async def drop_table(
         self,
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         ignore_missing: bool = False,
     ):
         """Drop a table from the database.
@@ -1683,34 +1687,34 @@ class AsyncConnection(object):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to drop the table from.
             Empty list represents root namespace.
         ignore_missing: bool, default False
             If True, ignore if the table does not exist.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         try:
-            await self._inner.drop_table(name, namespace=namespace)
+            await self._inner.drop_table(name, namespace_path=namespace_path)
         except ValueError as e:
             if not ignore_missing:
                 raise e
             if f"Table '{name}' was not found" not in str(e):
                 raise e
 
-    async def drop_all_tables(self, namespace: Optional[List[str]] = None):
+    async def drop_all_tables(self, namespace_path: Optional[List[str]] = None):
         """Drop all tables from the database.
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to drop all tables from.
             None or empty list represents root namespace.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
-        await self._inner.drop_all_tables(namespace=namespace)
+        await self._inner.drop_all_tables(namespace_path=namespace_path)
 
     @deprecation.deprecated(
         deprecated_in="0.15.1",

python/python/lancedb/embeddings/utils.py

@@ -10,6 +10,7 @@ import sys
 import threading
 import time
 import urllib.error
+import urllib.request
 import weakref
 import logging
 from functools import wraps

python/python/lancedb/expr.py

@@ -0,0 +1,298 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+"""Type-safe expression builder for filters and projections.
+
+Instead of writing raw SQL strings you can build expressions with Python
+operators::
+
+    from lancedb.expr import col, lit
+
+    # filter: age > 18 AND status = 'active'
+    filt = (col("age") > lit(18)) & (col("status") == lit("active"))
+
+    # projection: compute a derived column
+    proj = {"score": col("raw_score") * lit(1.5)}
+
+    table.search().where(filt).select(proj).to_list()
+"""
+
+from __future__ import annotations
+
+from typing import Union
+
+import pyarrow as pa
+
+from lancedb._lancedb import PyExpr, expr_col, expr_lit, expr_func
+
+__all__ = ["Expr", "col", "lit", "func"]
+
+_STR_TO_PA_TYPE: dict = {
+    "bool": pa.bool_(),
+    "boolean": pa.bool_(),
+    "int8": pa.int8(),
+    "int16": pa.int16(),
+    "int32": pa.int32(),
+    "int64": pa.int64(),
+    "uint8": pa.uint8(),
+    "uint16": pa.uint16(),
+    "uint32": pa.uint32(),
+    "uint64": pa.uint64(),
+    "float16": pa.float16(),
+    "float32": pa.float32(),
+    "float": pa.float32(),
+    "float64": pa.float64(),
+    "double": pa.float64(),
+    "string": pa.string(),
+    "utf8": pa.string(),
+    "str": pa.string(),
+    "large_string": pa.large_utf8(),
+    "large_utf8": pa.large_utf8(),
+    "date32": pa.date32(),
+    "date": pa.date32(),
+    "date64": pa.date64(),
+}
+
+
+def _coerce(value: "ExprLike") -> "Expr":
+    """Return *value* as an :class:`Expr`, wrapping plain Python values via
+    :func:`lit` if needed."""
+    if isinstance(value, Expr):
+        return value
+    return lit(value)
+
+
+# Type alias used in annotations.
+ExprLike = Union["Expr", bool, int, float, str]
+
+
+class Expr:
+    """A type-safe expression node.
+
+    Construct instances with :func:`col` and :func:`lit`, then combine them
+    using Python operators or the named methods below.
+
+    Examples
+    --------
+    >>> from lancedb.expr import col, lit
+    >>> filt = (col("age") > lit(18)) & (col("name").lower() == lit("alice"))
+    >>> proj = {"double": col("x") * lit(2)}
+    """
+
+    # Make Expr unhashable so that == returns an Expr rather than being used
+    # for dict keys / set membership.
+    __hash__ = None  # type: ignore[assignment]
+
+    def __init__(self, inner: PyExpr) -> None:
+        self._inner = inner
+
+    # ── comparisons ──────────────────────────────────────────────────────────
+
+    def __eq__(self, other: ExprLike) -> "Expr":  # type: ignore[override]
+        """Equal to (``col("x") == 1``)."""
+        return Expr(self._inner.eq(_coerce(other)._inner))
+
+    def __ne__(self, other: ExprLike) -> "Expr":  # type: ignore[override]
+        """Not equal to (``col("x") != 1``)."""
+        return Expr(self._inner.ne(_coerce(other)._inner))
+
+    def __lt__(self, other: ExprLike) -> "Expr":
+        """Less than (``col("x") < 1``)."""
+        return Expr(self._inner.lt(_coerce(other)._inner))
+
+    def __le__(self, other: ExprLike) -> "Expr":
+        """Less than or equal to (``col("x") <= 1``)."""
+        return Expr(self._inner.lte(_coerce(other)._inner))
+
+    def __gt__(self, other: ExprLike) -> "Expr":
+        """Greater than (``col("x") > 1``)."""
+        return Expr(self._inner.gt(_coerce(other)._inner))
+
+    def __ge__(self, other: ExprLike) -> "Expr":
+        """Greater than or equal to (``col("x") >= 1``)."""
+        return Expr(self._inner.gte(_coerce(other)._inner))
+
+    # ── logical ──────────────────────────────────────────────────────────────
+
+    def __and__(self, other: "Expr") -> "Expr":
+        """Logical AND (``expr_a & expr_b``)."""
+        return Expr(self._inner.and_(_coerce(other)._inner))
+
+    def __or__(self, other: "Expr") -> "Expr":
+        """Logical OR (``expr_a | expr_b``)."""
+        return Expr(self._inner.or_(_coerce(other)._inner))
+
+    def __invert__(self) -> "Expr":
+        """Logical NOT (``~expr``)."""
+        return Expr(self._inner.not_())
+
+    # ── arithmetic ───────────────────────────────────────────────────────────
+
+    def __add__(self, other: ExprLike) -> "Expr":
+        """Add (``col("x") + 1``)."""
+        return Expr(self._inner.add(_coerce(other)._inner))
+
+    def __radd__(self, other: ExprLike) -> "Expr":
+        """Right-hand add (``1 + col("x")``)."""
+        return Expr(_coerce(other)._inner.add(self._inner))
+
+    def __sub__(self, other: ExprLike) -> "Expr":
+        """Subtract (``col("x") - 1``)."""
+        return Expr(self._inner.sub(_coerce(other)._inner))
+
+    def __rsub__(self, other: ExprLike) -> "Expr":
+        """Right-hand subtract (``1 - col("x")``)."""
+        return Expr(_coerce(other)._inner.sub(self._inner))
+
+    def __mul__(self, other: ExprLike) -> "Expr":
+        """Multiply (``col("x") * 2``)."""
+        return Expr(self._inner.mul(_coerce(other)._inner))
+
+    def __rmul__(self, other: ExprLike) -> "Expr":
+        """Right-hand multiply (``2 * col("x")``)."""
+        return Expr(_coerce(other)._inner.mul(self._inner))
+
+    def __truediv__(self, other: ExprLike) -> "Expr":
+        """Divide (``col("x") / 2``)."""
+        return Expr(self._inner.div(_coerce(other)._inner))
+
+    def __rtruediv__(self, other: ExprLike) -> "Expr":
+        """Right-hand divide (``1 / col("x")``)."""
+        return Expr(_coerce(other)._inner.div(self._inner))
+
+    # ── string methods ───────────────────────────────────────────────────────
+
+    def lower(self) -> "Expr":
+        """Convert string column values to lowercase."""
+        return Expr(self._inner.lower())
+
+    def upper(self) -> "Expr":
+        """Convert string column values to uppercase."""
+        return Expr(self._inner.upper())
+
+    def contains(self, substr: "ExprLike") -> "Expr":
+        """Return True where the string contains *substr*."""
+        return Expr(self._inner.contains(_coerce(substr)._inner))
+
+    # ── type cast ────────────────────────────────────────────────────────────
+
+    def cast(self, data_type: Union[str, "pa.DataType"]) -> "Expr":
+        """Cast values to *data_type*.
+
+        Parameters
+        ----------
+        data_type:
+            A PyArrow ``DataType`` (e.g. ``pa.int32()``) or one of the type
+            name strings: ``"bool"``, ``"int8"``, ``"int16"``, ``"int32"``,
+            ``"int64"``, ``"uint8"``–``"uint64"``, ``"float32"``,
+            ``"float64"``, ``"string"``, ``"date32"``, ``"date64"``.
+        """
+        if isinstance(data_type, str):
+            try:
+                data_type = _STR_TO_PA_TYPE[data_type]
+            except KeyError:
+                raise ValueError(
+                    f"unsupported data type: '{data_type}'. Supported: "
+                    f"{', '.join(_STR_TO_PA_TYPE)}"
+                )
+        return Expr(self._inner.cast(data_type))
+
+    # ── named comparison helpers (alternative to operators) ──────────────────
+
+    def eq(self, other: ExprLike) -> "Expr":
+        """Equal to."""
+        return self.__eq__(other)
+
+    def ne(self, other: ExprLike) -> "Expr":
+        """Not equal to."""
+        return self.__ne__(other)
+
+    def lt(self, other: ExprLike) -> "Expr":
+        """Less than."""
+        return self.__lt__(other)
+
+    def lte(self, other: ExprLike) -> "Expr":
+        """Less than or equal to."""
+        return self.__le__(other)
+
+    def gt(self, other: ExprLike) -> "Expr":
+        """Greater than."""
+        return self.__gt__(other)
+
+    def gte(self, other: ExprLike) -> "Expr":
+        """Greater than or equal to."""
+        return self.__ge__(other)
+
+    def and_(self, other: "Expr") -> "Expr":
+        """Logical AND."""
+        return self.__and__(other)
+
+    def or_(self, other: "Expr") -> "Expr":
+        """Logical OR."""
+        return self.__or__(other)
+
+    # ── utilities ────────────────────────────────────────────────────────────
+
+    def to_sql(self) -> str:
+        """Render the expression as a SQL string (useful for debugging)."""
+        return self._inner.to_sql()
+
+    def __repr__(self) -> str:
+        return f"Expr({self._inner.to_sql()})"
+
+
+# ── free functions ────────────────────────────────────────────────────────────
+
+
+def col(name: str) -> Expr:
+    """Reference a table column by name.
+
+    Parameters
+    ----------
+    name:
+        The column name.
+
+    Examples
+    --------
+    >>> from lancedb.expr import col, lit
+    >>> col("age") > lit(18)
+    Expr((age > 18))
+    """
+    return Expr(expr_col(name))
+
+
+def lit(value: Union[bool, int, float, str]) -> Expr:
+    """Create a literal (constant) value expression.
+
+    Parameters
+    ----------
+    value:
+        A Python ``bool``, ``int``, ``float``, or ``str``.
+
+    Examples
+    --------
+    >>> from lancedb.expr import col, lit
+    >>> col("price") * lit(1.1)
+    Expr((price * 1.1))
+    """
+    return Expr(expr_lit(value))
+
+
+def func(name: str, *args: ExprLike) -> Expr:
+    """Call an arbitrary SQL function by name.
+
+    Parameters
+    ----------
+    name:
+        The SQL function name (e.g. ``"lower"``, ``"upper"``).
+    *args:
+        The function arguments as :class:`Expr` or plain Python literals.
+
+    Examples
+    --------
+    >>> from lancedb.expr import col, func
+    >>> func("lower", col("name"))
+    Expr(lower(name))
+    """
+    inner_args = [_coerce(a)._inner for a in args]
+    return Expr(expr_func(name, inner_args))

python/python/lancedb/io.py

@@ -2,70 +2,3 @@
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 """I/O utilities and interfaces for LanceDB."""
-
-from abc import ABC, abstractmethod
-from typing import Dict
-
-
-class StorageOptionsProvider(ABC):
-    """Abstract base class for providing storage options to LanceDB tables.
-
-    Storage options providers enable automatic credential refresh for cloud
-    storage backends (e.g., AWS S3, Azure Blob Storage, GCS). When credentials
-    have an expiration time, the provider's fetch_storage_options() method will
-    be called periodically to get fresh credentials before they expire.
-
-    Example
-    -------
-    >>> class MyProvider(StorageOptionsProvider):
-    ...     def fetch_storage_options(self) -> Dict[str, str]:
-    ...         # Fetch fresh credentials from your credential manager
-    ...         return {
-    ...             "aws_access_key_id": "...",
-    ...             "aws_secret_access_key": "...",
-    ...             "expires_at_millis": "1234567890000"  # Optional
-    ...         }
-    """
-
-    @abstractmethod
-    def fetch_storage_options(self) -> Dict[str, str]:
-        """Fetch fresh storage credentials.
-
-        This method is called by LanceDB when credentials need to be refreshed.
-        If the returned dictionary contains an "expires_at_millis" key with a
-        Unix timestamp in milliseconds, LanceDB will automatically refresh the
-        credentials before that time. If the key is not present, credentials
-        are assumed to not expire.
-
-        Returns
-        -------
-        Dict[str, str]
-            Dictionary containing cloud storage credentials and optionally an
-            expiration time:
-            - "expires_at_millis" (optional): Unix timestamp in milliseconds when
-              credentials expire
-            - Provider-specific credential keys (e.g., aws_access_key_id,
-              aws_secret_access_key, etc.)
-
-        Raises
-        ------
-        RuntimeError
-            If credentials cannot be fetched or are invalid
-        """
-        pass
-
-    def provider_id(self) -> str:
-        """Return a human-readable unique identifier for this provider instance.
-
-        This identifier is used for caching and equality comparison. Two providers
-        with the same ID will share the same cached object store connection.
-
-        The default implementation uses the class name and string representation.
-        Override this method if you need custom identification logic.
-
-        Returns
-        -------
-        str
-            A unique identifier for this provider instance
-        """
-        return f"{self.__class__.__name__} {{ repr: {str(self)!r} }}"

python/python/lancedb/query.py

@@ -38,6 +38,7 @@ from .rerankers.base import Reranker
 from .rerankers.rrf import RRFReranker
 from .rerankers.util import check_reranker_result
 from .util import flatten_columns
+from .expr import Expr
 from lancedb._lancedb import fts_query_to_json
 from typing_extensions import Annotated
 
@@ -70,7 +71,7 @@ def ensure_vector_query(
 ) -> Union[List[float], List[List[float]], pa.Array, List[pa.Array]]:
     if isinstance(val, list):
         if len(val) == 0:
-            return ValueError("Vector query must be a non-empty list")
+            raise ValueError("Vector query must be a non-empty list")
         sample = val[0]
     else:
         if isinstance(val, float):
@@ -83,7 +84,7 @@ def ensure_vector_query(
         return val
     if isinstance(sample, list):
         if len(sample) == 0:
-            return ValueError("Vector query must be a non-empty list")
+            raise ValueError("Vector query must be a non-empty list")
         if isinstance(sample[0], float):
             # val is list of list of floats
             return val
@@ -449,8 +450,8 @@ class Query(pydantic.BaseModel):
         ensure_vector_query,
     ] = None
 
-    # sql filter to refine the query with
+    # sql filter or type-safe Expr to refine the query with
-    filter: Optional[str] = None
+    filter: Optional[Union[str, Expr]] = None
 
     # if True then apply the filter after vector search
     postfilter: Optional[bool] = None
@@ -464,8 +465,8 @@ class Query(pydantic.BaseModel):
     # distance type to use for vector search
     distance_type: Optional[str] = None
 
-    # which columns to return in the results
+    # which columns to return in the results (dict values may be str or Expr)
-    columns: Optional[Union[List[str], Dict[str, str]]] = None
+    columns: Optional[Union[List[str], Dict[str, Union[str, Expr]]]] = None
 
     # minimum number of IVF partitions to search
     #
@@ -606,6 +607,7 @@ class LanceQueryBuilder(ABC):
                 query,
                 ordering_field_name=ordering_field_name,
                 fts_columns=fts_columns,
+                fast_search=fast_search,
             )
 
         if isinstance(query, list):
@@ -855,14 +857,15 @@ class LanceQueryBuilder(ABC):
             self._offset = offset
         return self
 
-    def select(self, columns: Union[list[str], dict[str, str]]) -> Self:
+    def select(self, columns: Union[list[str], dict[str, Union[str, Expr]]]) -> Self:
         """Set the columns to return.
 
         Parameters
         ----------
-        columns: list of str, or dict of str to str default None
+        columns: list of str, or dict of str to str or Expr
             List of column names to be fetched.
-            Or a dictionary of column names to SQL expressions.
+            Or a dictionary of column names to SQL expressions or
+            :class:`~lancedb.expr.Expr` objects.
             All columns are fetched if None or unspecified.
 
         Returns
@@ -876,15 +879,15 @@ class LanceQueryBuilder(ABC):
             raise ValueError("columns must be a list or a dictionary")
         return self
 
-    def where(self, where: str, prefilter: bool = True) -> Self:
+    def where(self, where: Union[str, Expr], prefilter: bool = True) -> Self:
         """Set the where clause.
 
         Parameters
         ----------
-        where: str
+        where: str or :class:`~lancedb.expr.Expr`
-            The where clause which is a valid SQL where clause. See
+            The filter condition.  Can be a SQL string or a type-safe
-            `Lance filter pushdown <https://lance.org/guide/read_and_write#filter-push-down>`_
+            :class:`~lancedb.expr.Expr` built with :func:`~lancedb.expr.col`
-            for valid SQL expressions.
+            and :func:`~lancedb.expr.lit`.
         prefilter: bool, default True
             If True, apply the filter before vector search, otherwise the
             filter is applied on the result of vector search.
@@ -1354,15 +1357,17 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
 
         return result_set
 
-    def where(self, where: str, prefilter: bool = None) -> LanceVectorQueryBuilder:
+    def where(
+        self, where: Union[str, Expr], prefilter: bool = None
+    ) -> LanceVectorQueryBuilder:
         """Set the where clause.
 
         Parameters
         ----------
-        where: str
+        where: str or :class:`~lancedb.expr.Expr`
-            The where clause which is a valid SQL where clause. See
+            The filter condition.  Can be a SQL string or a type-safe
-            `Lance filter pushdown <https://lance.org/guide/read_and_write#filter-push-down>`_
+            :class:`~lancedb.expr.Expr` built with :func:`~lancedb.expr.col`
-            for valid SQL expressions.
+            and :func:`~lancedb.expr.lit`.
         prefilter: bool, default True
             If True, apply the filter before vector search, otherwise the
             filter is applied on the result of vector search.
@@ -1456,12 +1461,14 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
         query: str | FullTextQuery,
         ordering_field_name: Optional[str] = None,
         fts_columns: Optional[Union[str, List[str]]] = None,
+        fast_search: bool = None,
     ):
         super().__init__(table)
         self._query = query
         self._phrase_query = False
         self.ordering_field_name = ordering_field_name
         self._reranker = None
+        self._fast_search = fast_search
         if isinstance(fts_columns, str):
             fts_columns = [fts_columns]
         self._fts_columns = fts_columns
@@ -1483,6 +1490,19 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
         self._phrase_query = phrase_query
         return self
 
+    def fast_search(self) -> LanceFtsQueryBuilder:
+        """
+        Skip a flat search of unindexed data. This will improve
+        search performance but search results will not include unindexed data.
+
+        Returns
+        -------
+        LanceFtsQueryBuilder
+            The LanceFtsQueryBuilder object.
+        """
+        self._fast_search = True
+        return self
+
     def to_query_object(self) -> Query:
         return Query(
             columns=self._columns,
@@ -1494,6 +1514,7 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
                 query=self._query, columns=self._fts_columns
             ),
             offset=self._offset,
+            fast_search=self._fast_search,
         )
 
     def output_schema(self) -> pa.Schema:
@@ -2188,8 +2209,8 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
             self._vector_query.select(self._columns)
             self._fts_query.select(self._columns)
         if self._where:
-            self._vector_query.where(self._where, self._postfilter)
+            self._vector_query.where(self._where, not self._postfilter)
-            self._fts_query.where(self._where, self._postfilter)
+            self._fts_query.where(self._where, not self._postfilter)
         if self._with_row_id:
             self._vector_query.with_row_id(True)
             self._fts_query.with_row_id(True)
@@ -2269,10 +2290,20 @@ class AsyncQueryBase(object):
         """
         if isinstance(columns, list) and all(isinstance(c, str) for c in columns):
             self._inner.select_columns(columns)
-        elif isinstance(columns, dict) and all(
+        elif isinstance(columns, dict) and all(isinstance(k, str) for k in columns):
-            isinstance(k, str) and isinstance(v, str) for k, v in columns.items()
+            if any(isinstance(v, Expr) for v in columns.values()):
-        ):
+                # At least one value is an Expr — use the type-safe path.
-            self._inner.select(list(columns.items()))
+                from .expr import _coerce
+
+                pairs = [(k, _coerce(v)._inner) for k, v in columns.items()]
+                self._inner.select_expr(pairs)
+            elif all(isinstance(v, str) for v in columns.values()):
+                self._inner.select(list(columns.items()))
+            else:
+                raise TypeError(
+                    "dict values must be str or Expr, got "
+                    + str({k: type(v) for k, v in columns.items()})
+                )
         else:
             raise TypeError("columns must be a list of column names or a dict")
         return self
@@ -2512,11 +2543,13 @@ class AsyncStandardQuery(AsyncQueryBase):
         """
         super().__init__(inner)
 
-    def where(self, predicate: str) -> Self:
+    def where(self, predicate: Union[str, Expr]) -> Self:
         """
         Only return rows matching the given predicate
 
-        The predicate should be supplied as an SQL query string.
+        The predicate can be a SQL string or a type-safe
+        :class:`~lancedb.expr.Expr` built with :func:`~lancedb.expr.col`
+        and :func:`~lancedb.expr.lit`.
 
         Examples
         --------
@@ -2528,7 +2561,10 @@ class AsyncStandardQuery(AsyncQueryBase):
         Filtering performance can often be improved by creating a scalar index
         on the filter column(s).
         """
-        self._inner.where(predicate)
+        if isinstance(predicate, Expr):
+            self._inner.where_expr(predicate._inner)
+        else:
+            self._inner.where(predicate)
         return self
 
     def limit(self, limit: int) -> Self:

python/python/lancedb/remote/db.py

@@ -111,7 +111,7 @@ class RemoteDBConnection(DBConnection):
     @override
     def list_namespaces(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListNamespacesResponse:
@@ -119,7 +119,7 @@ class RemoteDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The parent namespace to list namespaces in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -133,18 +133,18 @@ class RemoteDBConnection(DBConnection):
         ListNamespacesResponse
             Response containing namespace names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.list_namespaces(
-                namespace=namespace, page_token=page_token, limit=limit
+                namespace_path=namespace_path, page_token=page_token, limit=limit
             )
         )
 
     @override
     def create_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         properties: Optional[Dict[str, str]] = None,
     ) -> CreateNamespaceResponse:
@@ -152,7 +152,7 @@ class RemoteDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to create.
         mode: str, optional
             Creation mode - "create" (fail if exists), "exist_ok" (skip if exists),
@@ -167,14 +167,14 @@ class RemoteDBConnection(DBConnection):
         """
         return LOOP.run(
             self._conn.create_namespace(
-                namespace=namespace, mode=mode, properties=properties
+                namespace_path=namespace_path, mode=mode, properties=properties
             )
         )
 
     @override
     def drop_namespace(
         self,
-        namespace: List[str],
+        namespace_path: List[str],
         mode: Optional[str] = None,
         behavior: Optional[str] = None,
     ) -> DropNamespaceResponse:
@@ -182,7 +182,7 @@ class RemoteDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to drop.
         mode: str, optional
             Whether to skip if not exists ("SKIP") or fail ("FAIL"). Case insensitive.
@@ -196,16 +196,20 @@ class RemoteDBConnection(DBConnection):
             Response containing properties and transaction_id if applicable.
         """
         return LOOP.run(
-            self._conn.drop_namespace(namespace=namespace, mode=mode, behavior=behavior)
+            self._conn.drop_namespace(
+                namespace_path=namespace_path, mode=mode, behavior=behavior
+            )
         )
 
     @override
-    def describe_namespace(self, namespace: List[str]) -> DescribeNamespaceResponse:
+    def describe_namespace(
+        self, namespace_path: List[str]
+    ) -> DescribeNamespaceResponse:
         """Describe a namespace.
 
         Parameters
         ----------
-        namespace: List[str]
+        namespace_path: List[str]
             The namespace identifier to describe.
 
         Returns
@@ -213,12 +217,12 @@ class RemoteDBConnection(DBConnection):
         DescribeNamespaceResponse
             Response containing the namespace properties.
         """
-        return LOOP.run(self._conn.describe_namespace(namespace=namespace))
+        return LOOP.run(self._conn.describe_namespace(namespace_path=namespace_path))
 
     @override
     def list_tables(
         self,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         page_token: Optional[str] = None,
         limit: Optional[int] = None,
     ) -> ListTablesResponse:
@@ -226,7 +230,7 @@ class RemoteDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to list tables in.
             None or empty list represents root namespace.
         page_token: str, optional
@@ -240,11 +244,11 @@ class RemoteDBConnection(DBConnection):
         ListTablesResponse
             Response containing table names and optional page_token for pagination.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.list_tables(
-                namespace=namespace, page_token=page_token, limit=limit
+                namespace_path=namespace_path, page_token=page_token, limit=limit
             )
         )
 
@@ -254,7 +258,7 @@ class RemoteDBConnection(DBConnection):
         page_token: Optional[str] = None,
         limit: int = 10,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
     ) -> Iterable[str]:
         """List the names of all tables in the database.
 
@@ -263,7 +267,7 @@ class RemoteDBConnection(DBConnection):
 
         Parameters
         ----------
-        namespace: List[str], default []
+        namespace_path: List[str], default []
             The namespace to list tables in.
             Empty list represents root namespace.
         page_token: str
@@ -282,11 +286,11 @@ class RemoteDBConnection(DBConnection):
             DeprecationWarning,
             stacklevel=2,
         )
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         return LOOP.run(
             self._conn.table_names(
-                namespace=namespace, start_after=page_token, limit=limit
+                namespace_path=namespace_path, start_after=page_token, limit=limit
             )
         )
 
@@ -295,7 +299,7 @@ class RemoteDBConnection(DBConnection):
         self,
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
         index_cache_size: Optional[int] = None,
     ) -> Table:
@@ -305,7 +309,7 @@ class RemoteDBConnection(DBConnection):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to open the table from.
             None or empty list represents root namespace.
 
@@ -315,15 +319,15 @@ class RemoteDBConnection(DBConnection):
         """
         from .table import RemoteTable
 
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         if index_cache_size is not None:
             logging.info(
                 "index_cache_size is ignored in LanceDb Cloud"
                 " (there is no local cache to configure)"
             )
 
-        table = LOOP.run(self._conn.open_table(name, namespace=namespace))
+        table = LOOP.run(self._conn.open_table(name, namespace_path=namespace_path))
         return RemoteTable(table, self.db_name)
 
     def clone_table(
@@ -331,7 +335,7 @@ class RemoteDBConnection(DBConnection):
         target_table_name: str,
         source_uri: str,
         *,
-        target_namespace: Optional[List[str]] = None,
+        target_namespace_path: Optional[List[str]] = None,
         source_version: Optional[int] = None,
         source_tag: Optional[str] = None,
         is_shallow: bool = True,
@@ -344,7 +348,7 @@ class RemoteDBConnection(DBConnection):
             The name of the target table to create.
         source_uri: str
             The URI of the source table to clone from.
-        target_namespace: List[str], optional
+        target_namespace_path: List[str], optional
             The namespace for the target table.
             None or empty list represents root namespace.
         source_version: int, optional
@@ -361,13 +365,13 @@ class RemoteDBConnection(DBConnection):
         """
         from .table import RemoteTable
 
-        if target_namespace is None:
+        if target_namespace_path is None:
-            target_namespace = []
+            target_namespace_path = []
         table = LOOP.run(
             self._conn.clone_table(
                 target_table_name,
                 source_uri,
-                target_namespace=target_namespace,
+                target_namespace_path=target_namespace_path,
                 source_version=source_version,
                 source_tag=source_tag,
                 is_shallow=is_shallow,
@@ -387,7 +391,7 @@ class RemoteDBConnection(DBConnection):
         exist_ok: bool = False,
         embedding_functions: Optional[List[EmbeddingFunctionConfig]] = None,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
     ) -> Table:
         """Create a [Table][lancedb.table.Table] in the database.
 
@@ -395,7 +399,7 @@ class RemoteDBConnection(DBConnection):
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to create the table in.
             None or empty list represents root namespace.
         data: The data to initialize the table, *optional*
@@ -495,8 +499,8 @@ class RemoteDBConnection(DBConnection):
                 mode = "exist_ok"
             elif not mode:
                 mode = "exist_ok"
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         validate_table_name(name)
         if embedding_functions is not None:
             logging.warning(
@@ -511,7 +515,7 @@ class RemoteDBConnection(DBConnection):
             self._conn.create_table(
                 name,
                 data,
-                namespace=namespace,
+                namespace_path=namespace_path,
                 mode=mode,
                 schema=schema,
                 on_bad_vectors=on_bad_vectors,
@@ -521,28 +525,28 @@ class RemoteDBConnection(DBConnection):
         return RemoteTable(table, self.db_name)
 
     @override
-    def drop_table(self, name: str, namespace: Optional[List[str]] = None):
+    def drop_table(self, name: str, namespace_path: Optional[List[str]] = None):
         """Drop a table from the database.
 
         Parameters
         ----------
         name: str
             The name of the table.
-        namespace: List[str], optional
+        namespace_path: List[str], optional
             The namespace to drop the table from.
             None or empty list represents root namespace.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
-        LOOP.run(self._conn.drop_table(name, namespace=namespace))
+        LOOP.run(self._conn.drop_table(name, namespace_path=namespace_path))
 
     @override
     def rename_table(
         self,
         cur_name: str,
         new_name: str,
-        cur_namespace: Optional[List[str]] = None,
+        cur_namespace_path: Optional[List[str]] = None,
-        new_namespace: Optional[List[str]] = None,
+        new_namespace_path: Optional[List[str]] = None,
     ):
         """Rename a table in the database.
 
@@ -553,19 +557,19 @@ class RemoteDBConnection(DBConnection):
         new_name: str
             The new name of the table.
         """
-        if cur_namespace is None:
+        if cur_namespace_path is None:
-            cur_namespace = []
+            cur_namespace_path = []
-        if new_namespace is None:
+        if new_namespace_path is None:
-            new_namespace = []
+            new_namespace_path = []
         LOOP.run(
             self._conn.rename_table(
                 cur_name,
                 new_name,
-                cur_namespace=cur_namespace,
+                cur_namespace_path=cur_namespace_path,
-                new_namespace=new_namespace,
+                new_namespace_path=new_namespace_path,
             )
         )
 
     async def close(self):
         """Close the connection to the database."""
-        self._client.close()
+        self._conn.close()

python/python/lancedb/remote/table.py

@@ -4,7 +4,7 @@
 from datetime import timedelta
 import logging
 from functools import cached_property
-from typing import Dict, Iterable, List, Optional, Union, Literal
+from typing import Any, Callable, Dict, Iterable, List, Optional, Union, Literal
 import warnings
 
 from lancedb._lancedb import (
@@ -35,6 +35,7 @@ import pyarrow as pa
 from lancedb.common import DATA, VEC, VECTOR_COLUMN_NAME
 from lancedb.merge import LanceMergeInsertBuilder
 from lancedb.embeddings import EmbeddingFunctionRegistry
+from lancedb.table import _normalize_progress
 
 from ..query import LanceVectorQueryBuilder, LanceQueryBuilder, LanceTakeQueryBuilder
 from ..table import AsyncTable, IndexStatistics, Query, Table, Tags
@@ -218,8 +219,6 @@ class RemoteTable(Table):
         train: bool = True,
     ):
         """Create an index on the table.
-        Currently, the only parameters that matter are
-        the metric and the vector column name.
 
         Parameters
         ----------
@@ -250,11 +249,6 @@ class RemoteTable(Table):
         >>> table.create_index("l2", "vector") # doctest: +SKIP
         """
 
-        if num_sub_vectors is not None:
-            logging.warning(
-                "num_sub_vectors is not supported on LanceDB cloud."
-                "This parameter will be tuned automatically."
-            )
         if accelerator is not None:
             logging.warning(
                 "GPU accelerator is not yet supported on LanceDB cloud."
@@ -315,6 +309,7 @@ class RemoteTable(Table):
         mode: str = "append",
         on_bad_vectors: str = "error",
         fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
     ) -> AddResult:
         """Add more data to the [Table](Table). It has the same API signature as
         the OSS version.
@@ -337,17 +332,29 @@ class RemoteTable(Table):
             One of "error", "drop", "fill".
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.
 
         Returns
         -------
         AddResult
             An object containing the new version number of the table after adding data.
         """
-        return LOOP.run(
+        progress, owns = _normalize_progress(progress)
-            self._table.add(
+        try:
-                data, mode=mode, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+            return LOOP.run(
+                self._table.add(
+                    data,
+                    mode=mode,
+                    on_bad_vectors=on_bad_vectors,
+                    fill_value=fill_value,
+                    progress=progress,
+                )
             )
-        )
+        finally:
+            if owns:
+                progress.close()
 
     def search(
         self,
@@ -647,6 +654,45 @@ class RemoteTable(Table):
     def drop_index(self, index_name: str):
         return LOOP.run(self._table.drop_index(index_name))
 
+    def prewarm_index(self, name: str) -> None:
+        """Prewarm an index in the table.
+
+        This is a hint to the database that the index will be accessed in the
+        future and should be loaded into memory if possible.  This can reduce
+        cold-start latency for subsequent queries.
+
+        This call initiates prewarming and returns once the request is accepted.
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        Parameters
+        ----------
+        name: str
+            The name of the index to prewarm
+        """
+        return LOOP.run(self._table.prewarm_index(name))
+
+    def prewarm_data(self, columns: Optional[List[str]] = None) -> None:
+        """Prewarm data for the table.
+
+        This is a hint to the database that the given columns will be accessed
+        in the future and the database should prefetch the data if possible.
+        Currently only supported on remote tables.
+
+        This call initiates prewarming and returns once the request is accepted.
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        This operation has a large upfront cost but can speed up future queries
+        that need to fetch the given columns.  Large columns such as embeddings
+        or binary data may not be practical to prewarm.  This feature is intended
+        for workloads that issue many queries against the same columns.
+
+        Parameters
+        ----------
+        columns: list of str, optional
+            The columns to prewarm. If None, all columns are prewarmed.
+        """
+        return LOOP.run(self._table.prewarm_data(columns))
+
     def wait_for_index(
         self, index_names: Iterable[str], timeout: timedelta = timedelta(seconds=300)
     ):

python/python/lancedb/scannable.py

@@ -0,0 +1,214 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+from dataclasses import dataclass
+from functools import singledispatch
+import sys
+from typing import Callable, Iterator, Optional
+from lancedb.arrow import to_arrow
+import pyarrow as pa
+import pyarrow.dataset as ds
+
+from .pydantic import LanceModel
+
+
+@dataclass
+class Scannable:
+    schema: pa.Schema
+    num_rows: Optional[int]
+    # Factory function to create a new reader each time (supports re-scanning)
+    reader: Callable[[], pa.RecordBatchReader]
+    # Whether reader can be called more than once. For example, an iterator can
+    # only be consumed once, while a DataFrame can be converted to a new reader
+    # each time.
+    rescannable: bool = True
+
+
+@singledispatch
+def to_scannable(data) -> Scannable:
+    # Fallback: try iterable protocol
+    if hasattr(data, "__iter__"):
+        return _from_iterable(iter(data))
+    raise NotImplementedError(f"to_scannable not implemented for type {type(data)}")
+
+
+@to_scannable.register(pa.RecordBatchReader)
+def _from_reader(data: pa.RecordBatchReader) -> Scannable:
+    # RecordBatchReader can only be consumed once - not rescannable
+    return Scannable(
+        schema=data.schema, num_rows=None, reader=lambda: data, rescannable=False
+    )
+
+
+@to_scannable.register(pa.RecordBatch)
+def _from_batch(data: pa.RecordBatch) -> Scannable:
+    return Scannable(
+        schema=data.schema,
+        num_rows=data.num_rows,
+        reader=lambda: pa.RecordBatchReader.from_batches(data.schema, [data]),
+    )
+
+
+@to_scannable.register(pa.Table)
+def _from_table(data: pa.Table) -> Scannable:
+    return Scannable(schema=data.schema, num_rows=data.num_rows, reader=data.to_reader)
+
+
+@to_scannable.register(ds.Dataset)
+def _from_dataset(data: ds.Dataset) -> Scannable:
+    return Scannable(
+        schema=data.schema,
+        num_rows=data.count_rows(),
+        reader=lambda: data.scanner().to_reader(),
+    )
+
+
+@to_scannable.register(ds.Scanner)
+def _from_scanner(data: ds.Scanner) -> Scannable:
+    # Scanner can only be consumed once - not rescannable
+    return Scannable(
+        schema=data.projected_schema,
+        num_rows=None,
+        reader=data.to_reader,
+        rescannable=False,
+    )
+
+
+@to_scannable.register(list)
+def _from_list(data: list) -> Scannable:
+    if not data:
+        raise ValueError("Cannot create table from empty list without a schema")
+    table = to_arrow(data)
+    return Scannable(
+        schema=table.schema, num_rows=table.num_rows, reader=table.to_reader
+    )
+
+
+@to_scannable.register(dict)
+def _from_dict(data: dict) -> Scannable:
+    raise ValueError("Cannot add a single dictionary to a table. Use a list.")
+
+
+@to_scannable.register(LanceModel)
+def _from_lance_model(data: LanceModel) -> Scannable:
+    raise ValueError("Cannot add a single LanceModel to a table. Use a list.")
+
+
+def _from_iterable(data: Iterator) -> Scannable:
+    first_item = next(data, None)
+    if first_item is None:
+        raise ValueError("Cannot create table from empty iterator")
+    first = to_arrow(first_item)
+    schema = first.schema
+
+    def iter():
+        yield from first.to_batches()
+        for item in data:
+            batch = to_arrow(item)
+            if batch.schema != schema:
+                try:
+                    batch = batch.cast(schema)
+                except pa.lib.ArrowInvalid:
+                    raise ValueError(
+                        f"Input iterator yielded a batch with schema that "
+                        f"does not match the schema of other batches.\n"
+                        f"Expected:\n{schema}\nGot:\n{batch.schema}"
+                    )
+            yield from batch.to_batches()
+
+    reader = pa.RecordBatchReader.from_batches(schema, iter())
+    return to_scannable(reader)
+
+
+_registered_modules: set[str] = set()
+
+
+def _register_optional_converters():
+    """Register converters for optional dependencies that are already imported."""
+
+    if "pandas" in sys.modules and "pandas" not in _registered_modules:
+        _registered_modules.add("pandas")
+        import pandas as pd
+
+        @to_arrow.register(pd.DataFrame)
+        def _arrow_from_pandas(data: pd.DataFrame) -> pa.Table:
+            table = pa.Table.from_pandas(data, preserve_index=False)
+            return table.replace_schema_metadata(None)
+
+        @to_scannable.register(pd.DataFrame)
+        def _from_pandas(data: pd.DataFrame) -> Scannable:
+            return to_scannable(_arrow_from_pandas(data))
+
+    if "polars" in sys.modules and "polars" not in _registered_modules:
+        _registered_modules.add("polars")
+        import polars as pl
+
+        @to_arrow.register(pl.DataFrame)
+        def _arrow_from_polars(data: pl.DataFrame) -> pa.Table:
+            return data.to_arrow()
+
+        @to_scannable.register(pl.DataFrame)
+        def _from_polars(data: pl.DataFrame) -> Scannable:
+            arrow = data.to_arrow()
+            return Scannable(
+                schema=arrow.schema, num_rows=len(data), reader=arrow.to_reader
+            )
+
+        @to_scannable.register(pl.LazyFrame)
+        def _from_polars_lazy(data: pl.LazyFrame) -> Scannable:
+            arrow = data.collect().to_arrow()
+            return Scannable(
+                schema=arrow.schema, num_rows=arrow.num_rows, reader=arrow.to_reader
+            )
+
+    if "datasets" in sys.modules and "datasets" not in _registered_modules:
+        _registered_modules.add("datasets")
+        from datasets import Dataset as HFDataset
+        from datasets import DatasetDict as HFDatasetDict
+
+        @to_scannable.register(HFDataset)
+        def _from_hf_dataset(data: HFDataset) -> Scannable:
+            table = data.data.table  # Access underlying Arrow table
+            return Scannable(
+                schema=table.schema, num_rows=len(data), reader=table.to_reader
+            )
+
+        @to_scannable.register(HFDatasetDict)
+        def _from_hf_dataset_dict(data: HFDatasetDict) -> Scannable:
+            # HuggingFace DatasetDict: combine all splits with a 'split' column
+            schema = data[list(data.keys())[0]].features.arrow_schema
+            if "split" not in schema.names:
+                schema = schema.append(pa.field("split", pa.string()))
+
+            def gen():
+                for split_name, dataset in data.items():
+                    for batch in dataset.data.to_batches():
+                        split_arr = pa.array(
+                            [split_name] * len(batch), type=pa.string()
+                        )
+                        yield pa.RecordBatch.from_arrays(
+                            list(batch.columns) + [split_arr], schema=schema
+                        )
+
+            total_rows = sum(len(dataset) for dataset in data.values())
+            return Scannable(
+                schema=schema,
+                num_rows=total_rows,
+                reader=lambda: pa.RecordBatchReader.from_batches(schema, gen()),
+            )
+
+    if "lance" in sys.modules and "lance" not in _registered_modules:
+        _registered_modules.add("lance")
+        import lance
+
+        @to_scannable.register(lance.LanceDataset)
+        def _from_lance(data: lance.LanceDataset) -> Scannable:
+            return Scannable(
+                schema=data.schema,
+                num_rows=data.count_rows(),
+                reader=lambda: data.scanner().to_reader(),
+            )
+
+
+# Register on module load
+_register_optional_converters()

python/python/lancedb/table.py

@@ -14,6 +14,7 @@ from functools import cached_property
 from typing import (
     TYPE_CHECKING,
     Any,
+    Callable,
     Dict,
     Iterable,
     List,
@@ -25,6 +26,8 @@ from typing import (
 )
 from urllib.parse import urlparse
 
+from lancedb.scannable import _register_optional_converters, to_scannable
+
 from . import __version__
 from lancedb.arrow import peek_reader
 from lancedb.background_loop import LOOP
@@ -86,7 +89,6 @@ from .index import lang_mapping
 
 if TYPE_CHECKING:
     from .db import LanceDBConnection
-    from .io import StorageOptionsProvider
     from ._lancedb import (
         Table as LanceDBTable,
         OptimizeStats,
@@ -275,7 +277,7 @@ def _sanitize_data(
 
     if metadata:
         new_metadata = target_schema.metadata or {}
-        new_metadata = new_metadata.update(metadata)
+        new_metadata.update(metadata)
         target_schema = target_schema.with_metadata(new_metadata)
 
     _validate_schema(target_schema)
@@ -554,6 +556,21 @@ def _table_uri(base: str, table_name: str) -> str:
     return join_uri(base, f"{table_name}.lance")
 
 
+def _normalize_progress(progress):
+    """Normalize a ``progress`` parameter for :meth:`Table.add`.
+
+    Returns ``(progress_obj, owns)`` where *owns* is True when we created a
+    tqdm bar that the caller must close.
+    """
+    if progress is True:
+        from tqdm.auto import tqdm
+
+        return tqdm(unit=" rows"), True
+    if progress is False or progress is None:
+        return None, False
+    return progress, False
+
+
 class Table(ABC):
     """
     A Table is a collection of Records in a LanceDB Database.
@@ -972,6 +989,7 @@ class Table(ABC):
         mode: AddMode = "append",
         on_bad_vectors: OnBadVectorsType = "error",
         fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
     ) -> AddResult:
         """Add more data to the [Table](Table).
 
@@ -993,6 +1011,29 @@ class Table(ABC):
             One of "error", "drop", "fill".
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            Progress reporting during the add operation. Can be:
+
+            - ``True`` to automatically create and display a tqdm progress
+              bar (requires ``tqdm`` to be installed)::
+
+                table.add(data, progress=True)
+
+            - A **callable** that receives a dict with keys ``output_rows``,
+              ``output_bytes``, ``total_rows``, ``elapsed_seconds``,
+              ``active_tasks``, ``total_tasks``, and ``done``::
+
+                def on_progress(p):
+                    print(f"{p['output_rows']}/{p['total_rows']} rows, "
+                          f"{p['active_tasks']}/{p['total_tasks']} workers")
+                table.add(data, progress=on_progress)
+
+            - A **tqdm-compatible** progress bar whose ``total`` and
+              ``update()`` will be called automatically. The postfix shows
+              write throughput (MB/s) and active worker count::
+
+                with tqdm() as pbar:
+                    table.add(data, progress=pbar)
 
         Returns
         -------
@@ -1329,7 +1370,7 @@ class Table(ABC):
         1  2  [3.0, 4.0]
         2  3  [5.0, 6.0]
         >>> table.delete("x = 2")
-        DeleteResult(version=2)
+        DeleteResult(num_deleted_rows=1, version=2)
         >>> table.to_pandas()
            x      vector
         0  1  [1.0, 2.0]
@@ -1343,7 +1384,7 @@ class Table(ABC):
         >>> to_remove
         '1, 5'
         >>> table.delete(f"x IN ({to_remove})")
-        DeleteResult(version=3)
+        DeleteResult(num_deleted_rows=1, version=3)
         >>> table.to_pandas()
            x      vector
         0  3  [5.0, 6.0]
@@ -1504,22 +1545,17 @@ class Table(ABC):
             in-progress operation (e.g. appending new data) and these files will not
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
         retrain: bool, default False
             This parameter is no longer used and is deprecated.
 
-        Experimental API
+        The frequency an application should call optimize is based on the frequency of
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
         data modifications.  If data is frequently added, deleted, or updated then
         optimize should be run frequently.  A good rule of thumb is to run optimize if
         you have added or modified 100,000 or more records or run more than 20 data
@@ -1739,29 +1775,34 @@ class LanceTable(Table):
         connection: "LanceDBConnection",
         name: str,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         index_cache_size: Optional[int] = None,
         location: Optional[str] = None,
+        namespace_client: Optional[Any] = None,
+        managed_versioning: Optional[bool] = None,
+        pushdown_operations: Optional[set] = None,
         _async: AsyncTable = None,
     ):
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         self._conn = connection
-        self._namespace = namespace
+        self._namespace_path = namespace_path
         self._location = location  # Store location for use in _dataset_path
+        self._namespace_client = namespace_client
+        self._pushdown_operations = pushdown_operations or set()
         if _async is not None:
             self._table = _async
         else:
             self._table = LOOP.run(
                 connection._conn.open_table(
                     name,
-                    namespace=namespace,
+                    namespace_path=namespace_path,
                     storage_options=storage_options,
-                    storage_options_provider=storage_options_provider,
                     index_cache_size=index_cache_size,
                     location=location,
+                    namespace_client=namespace_client,
+                    managed_versioning=managed_versioning,
                 )
             )
 
@@ -1772,13 +1813,13 @@ class LanceTable(Table):
     @property
     def namespace(self) -> List[str]:
         """Return the namespace path of the table."""
-        return self._namespace
+        return self._namespace_path
 
     @property
     def id(self) -> str:
         """Return the full identifier of the table (namespace$name)."""
-        if self._namespace:
+        if self._namespace_path:
-            return "$".join(self._namespace + [self.name])
+            return "$".join(self._namespace_path + [self.name])
         return self.name
 
     @classmethod
@@ -1799,22 +1840,26 @@ class LanceTable(Table):
         db,
         name,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         index_cache_size: Optional[int] = None,
         location: Optional[str] = None,
+        namespace_client: Optional[Any] = None,
+        managed_versioning: Optional[bool] = None,
+        pushdown_operations: Optional[set] = None,
     ):
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         tbl = cls(
             db,
             name,
-            namespace=namespace,
+            namespace_path=namespace_path,
             storage_options=storage_options,
-            storage_options_provider=storage_options_provider,
             index_cache_size=index_cache_size,
             location=location,
+            namespace_client=namespace_client,
+            managed_versioning=managed_versioning,
+            pushdown_operations=pushdown_operations,
         )
 
         # check the dataset exists
@@ -1846,6 +1891,16 @@ class LanceTable(Table):
                 "Please install with `pip install pylance`."
             )
 
+        if self._namespace_client is not None:
+            table_id = self._namespace_path + [self.name]
+            return lance.dataset(
+                version=self.version,
+                storage_options=self._conn.storage_options,
+                namespace_client=self._namespace_client,
+                table_id=table_id,
+                **kwargs,
+            )
+
         return lance.dataset(
             self._dataset_path,
             version=self.version,
@@ -2198,12 +2253,18 @@ class LanceTable(Table):
 
     def prewarm_index(self, name: str) -> None:
         """
-        Prewarms an index in the table
+        Prewarm an index in the table.
 
-        This loads the entire index into memory
+        This is a hint to the database that the index will be accessed in the
+        future and should be loaded into memory if possible.  This can reduce
+        cold-start latency for subsequent queries.
 
-        If the index does not fit into the available cache this call
+        This call initiates prewarming and returns once the request is accepted.
-        may be wasteful
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        It is generally wasteful to call this if the index does not fit into the
+        available cache.  Not all index types support prewarming; unsupported
+        indices will silently ignore the request.
 
         Parameters
         ----------
@@ -2212,6 +2273,29 @@ class LanceTable(Table):
         """
         return LOOP.run(self._table.prewarm_index(name))
 
+    def prewarm_data(self, columns: Optional[List[str]] = None) -> None:
+        """
+        Prewarm data for the table.
+
+        This is a hint to the database that the given columns will be accessed
+        in the future and the database should prefetch the data if possible.
+        Currently only supported on remote tables.
+
+        This call initiates prewarming and returns once the request is accepted.
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        This operation has a large upfront cost but can speed up future queries
+        that need to fetch the given columns.  Large columns such as embeddings
+        or binary data may not be practical to prewarm.  This feature is intended
+        for workloads that issue many queries against the same columns.
+
+        Parameters
+        ----------
+        columns: list of str, optional
+            The columns to prewarm. If None, all columns are prewarmed.
+        """
+        return LOOP.run(self._table.prewarm_data(columns))
+
     def wait_for_index(
         self, index_names: Iterable[str], timeout: timedelta = timedelta(seconds=300)
     ) -> None:
@@ -2447,6 +2531,7 @@ class LanceTable(Table):
         mode: AddMode = "append",
         on_bad_vectors: OnBadVectorsType = "error",
         fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
     ) -> AddResult:
         """Add data to the table.
         If vector columns are missing and the table
@@ -2465,17 +2550,29 @@ class LanceTable(Table):
             One of "error", "drop", "fill", "null".
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.
 
         Returns
         -------
         int
             The number of vectors in the table.
         """
-        return LOOP.run(
+        progress, owns = _normalize_progress(progress)
-            self._table.add(
+        try:
-                data, mode=mode, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+            return LOOP.run(
+                self._table.add(
+                    data,
+                    mode=mode,
+                    on_bad_vectors=on_bad_vectors,
+                    fill_value=fill_value,
+                    progress=progress,
+                )
             )
-        )
+        finally:
+            if owns:
+                progress.close()
 
     def merge(
         self,
@@ -2705,12 +2802,13 @@ class LanceTable(Table):
         fill_value: float = 0.0,
         embedding_functions: Optional[List[EmbeddingFunctionConfig]] = None,
         *,
-        namespace: Optional[List[str]] = None,
+        namespace_path: Optional[List[str]] = None,
         storage_options: Optional[Dict[str, str | bool]] = None,
-        storage_options_provider: Optional["StorageOptionsProvider"] = None,
         data_storage_version: Optional[str] = None,
         enable_v2_manifest_paths: Optional[bool] = None,
         location: Optional[str] = None,
+        namespace_client: Optional[Any] = None,
+        pushdown_operations: Optional[set] = None,
     ):
         """
         Create a new table.
@@ -2765,12 +2863,14 @@ class LanceTable(Table):
             Deprecated.  Set `storage_options` when connecting to the database and set
             `new_table_enable_v2_manifest_paths` in the options.
         """
-        if namespace is None:
+        if namespace_path is None:
-            namespace = []
+            namespace_path = []
         self = cls.__new__(cls)
         self._conn = db
-        self._namespace = namespace
+        self._namespace_path = namespace_path
         self._location = location
+        self._namespace_client = namespace_client
+        self._pushdown_operations = pushdown_operations or set()
 
         if data_storage_version is not None:
             warnings.warn(
@@ -2803,9 +2903,8 @@ class LanceTable(Table):
                 on_bad_vectors=on_bad_vectors,
                 fill_value=fill_value,
                 embedding_functions=embedding_functions,
-                namespace=namespace,
+                namespace_path=namespace_path,
                 storage_options=storage_options,
-                storage_options_provider=storage_options_provider,
                 location=location,
             )
         )
@@ -2874,6 +2973,15 @@ class LanceTable(Table):
         batch_size: Optional[int] = None,
         timeout: Optional[timedelta] = None,
     ) -> pa.RecordBatchReader:
+        if (
+            "QueryTable" in self._pushdown_operations
+            and self._namespace_client is not None
+        ):
+            from lancedb.namespace import _execute_server_side_query
+
+            table_id = self._namespace_path + [self.name]
+            return _execute_server_side_query(self._namespace_client, table_id, query)
+
         async_iter = LOOP.run(
             self._table._execute_query(query, batch_size=batch_size, timeout=timeout)
         )
@@ -2995,22 +3103,17 @@ class LanceTable(Table):
             in-progress operation (e.g. appending new data) and these files will not
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
         retrain: bool, default False
             This parameter is no longer used and is deprecated.
 
-        Experimental API
+        The frequency an application should call optimize is based on the frequency of
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
         data modifications.  If data is frequently added, deleted, or updated then
         optimize should be run frequently.  A good rule of thumb is to run optimize if
         you have added or modified 100,000 or more records or run more than 20 data
@@ -3611,19 +3714,47 @@ class AsyncTable:
         """
         Prewarm an index in the table.
 
+        This is a hint to the database that the index will be accessed in the
+        future and should be loaded into memory if possible.  This can reduce
+        cold-start latency for subsequent queries.
+
+        This call initiates prewarming and returns once the request is accepted.
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        It is generally wasteful to call this if the index does not fit into the
+        available cache.  Not all index types support prewarming; unsupported
+        indices will silently ignore the request.
+
         Parameters
         ----------
         name: str
             The name of the index to prewarm
-
-        Notes
-        -----
-        This will load the index into memory.  This may reduce the cold-start time for
-        future queries.  If the index does not fit in the cache then this call may be
-        wasteful.
         """
         await self._inner.prewarm_index(name)
 
+    async def prewarm_data(self, columns: Optional[List[str]] = None) -> None:
+        """
+        Prewarm data for the table.
+
+        This is a hint to the database that the given columns will be accessed
+        in the future and the database should prefetch the data if possible.
+        Currently only supported on remote tables.
+
+        This call initiates prewarming and returns once the request is accepted.
+        It is idempotent and safe to call from multiple clients concurrently.
+
+        This operation has a large upfront cost but can speed up future queries
+        that need to fetch the given columns.  Large columns such as embeddings
+        or binary data may not be practical to prewarm.  This feature is intended
+        for workloads that issue many queries against the same columns.
+
+        Parameters
+        ----------
+        columns: list of str, optional
+            The columns to prewarm. If None, all columns are prewarmed.
+        """
+        await self._inner.prewarm_data(columns)
+
     async def wait_for_index(
         self, index_names: Iterable[str], timeout: timedelta = timedelta(seconds=300)
     ) -> None:
@@ -3699,6 +3830,7 @@ class AsyncTable:
         mode: Optional[Literal["append", "overwrite"]] = "append",
         on_bad_vectors: Optional[OnBadVectorsType] = None,
         fill_value: Optional[float] = None,
+        progress: Optional[Union[bool, Callable, Any]] = None,
     ) -> AddResult:
         """Add more data to the [Table](Table).
 
@@ -3720,6 +3852,9 @@ class AsyncTable:
             One of "error", "drop", "fill", "null".
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: callable or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.
 
         """
         schema = await self.schema()
@@ -3727,18 +3862,41 @@ class AsyncTable:
             on_bad_vectors = "error"
         if fill_value is None:
             fill_value = 0.0
-        data = _sanitize_data(
-            data,
-            schema,
-            metadata=schema.metadata,
-            on_bad_vectors=on_bad_vectors,
-            fill_value=fill_value,
-            allow_subschema=True,
-        )
-        if isinstance(data, pa.Table):
-            data = data.to_reader()
 
-        return await self._inner.add(data, mode or "append")
+        # _santitize_data is an old code path, but we will use it until the
+        # new code path is ready.
+        if mode == "overwrite":
+            # For overwrite, apply the same preprocessing as create_table
+            # so vector columns are inferred as FixedSizeList.
+            data, _ = sanitize_create_table(
+                data, None, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+            )
+        elif on_bad_vectors != "error" or (
+            schema.metadata is not None and b"embedding_functions" in schema.metadata
+        ):
+            data = _sanitize_data(
+                data,
+                schema,
+                metadata=schema.metadata,
+                on_bad_vectors=on_bad_vectors,
+                fill_value=fill_value,
+                allow_subschema=True,
+            )
+        _register_optional_converters()
+        data = to_scannable(data)
+        progress, owns = _normalize_progress(progress)
+        try:
+            return await self._inner.add(data, mode or "append", progress=progress)
+        except RuntimeError as e:
+            if "Cast error" in str(e):
+                raise ValueError(e)
+            elif "Vector column contains NaN" in str(e):
+                raise ValueError(e)
+            else:
+                raise
+        finally:
+            if owns:
+                progress.close()
 
     def merge_insert(self, on: Union[str, Iterable[str]]) -> LanceMergeInsertBuilder:
         """
@@ -4061,7 +4219,7 @@ class AsyncTable:
             async_query = async_query.offset(query.offset)
         if query.columns:
             async_query = async_query.select(query.columns)
-        if query.filter:
+        if query.filter is not None:
             async_query = async_query.where(query.filter)
         if query.fast_search:
             async_query = async_query.fast_search()
@@ -4200,7 +4358,7 @@ class AsyncTable:
         1  2  [3.0, 4.0]
         2  3  [5.0, 6.0]
         >>> table.delete("x = 2")
-        DeleteResult(version=2)
+        DeleteResult(num_deleted_rows=1, version=2)
         >>> table.to_pandas()
            x      vector
         0  1  [1.0, 2.0]
@@ -4214,7 +4372,7 @@ class AsyncTable:
         >>> to_remove
         '1, 5'
         >>> table.delete(f"x IN ({to_remove})")
-        DeleteResult(version=3)
+        DeleteResult(num_deleted_rows=1, version=3)
         >>> table.to_pandas()
            x      vector
         0  3  [5.0, 6.0]
@@ -4537,22 +4695,17 @@ class AsyncTable:
             in-progress operation (e.g. appending new data) and these files will not
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
         retrain: bool, default False
             This parameter is no longer used and is deprecated.
 
-        Experimental API
+        The frequency an application should call optimize is based on the frequency of
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
         data modifications.  If data is frequently added, deleted, or updated then
         optimize should be run frequently.  A good rule of thumb is to run optimize if
         you have added or modified 100,000 or more records or run more than 20 data
@@ -4673,7 +4826,16 @@ class IndexStatistics:
     num_indexed_rows: int
     num_unindexed_rows: int
     index_type: Literal[
-        "IVF_PQ", "IVF_HNSW_PQ", "IVF_HNSW_SQ", "FTS", "BTREE", "BITMAP", "LABEL_LIST"
+        "IVF_FLAT",
+        "IVF_SQ",
+        "IVF_PQ",
+        "IVF_RQ",
+        "IVF_HNSW_SQ",
+        "IVF_HNSW_PQ",
+        "FTS",
+        "BTREE",
+        "BITMAP",
+        "LABEL_LIST",
     ]
     distance_type: Optional[Literal["l2", "cosine", "dot"]] = None
     num_indices: Optional[int] = None

python/python/lancedb/util.py

@@ -324,6 +324,16 @@ def _(value: list):
     return "[" + ", ".join(map(value_to_sql, value)) + "]"
 
 
+@value_to_sql.register(dict)
+def _(value: dict):
+    # https://datafusion.apache.org/user-guide/sql/scalar_functions.html#named-struct
+    return (
+        "named_struct("
+        + ", ".join(f"'{k}', {value_to_sql(v)}" for k, v in value.items())
+        + ")"
+    )
+
+
 @value_to_sql.register(np.ndarray)
 def _(value: np.ndarray):
     return value_to_sql(value.tolist())

python/python/tests/test_db.py

@@ -183,8 +183,8 @@ def test_table_names(tmp_db: lancedb.DBConnection):
     result = list(tmp_db.table_names("test2", limit=2))
     assert result == ["test3"], f"Expected ['test3'], got {result}"
 
-    # Test that namespace parameter can be passed as keyword
+    # Test that namespace_path parameter can be passed as keyword
-    result = list(tmp_db.table_names(namespace=[]))
+    result = list(tmp_db.table_names(namespace_path=[]))
     assert len(result) == 3
 
 
@@ -909,7 +909,7 @@ def test_local_namespace_operations(tmp_path):
         NotImplementedError,
         match="Namespace operations are not supported for listing database",
     ):
-        db.list_namespaces(namespace=["test"])
+        db.list_namespaces(namespace_path=["test"])
 
 
 def test_local_create_namespace_not_supported(tmp_path):

python/python/tests/test_embeddings.py

@@ -546,3 +546,24 @@ def test_openai_no_retry_on_401(mock_sleep):
     assert mock_func.call_count == 1
     # Verify that sleep was never called (no retries)
     assert mock_sleep.call_count == 0
+
+
+def test_url_retrieve_downloads_image():
+    """
+    Embedding functions like open-clip, siglip, and jinaai use url_retrieve()
+    to download images from HTTP URLs. For example, open_clip._to_pil() calls:
+
+        PIL_Image.open(io.BytesIO(url_retrieve(image)))
+
+    Verify that url_retrieve() can download an image and open it as PIL Image,
+    matching the real usage pattern in embedding functions.
+    """
+    import io
+
+    Image = pytest.importorskip("PIL.Image")
+    from lancedb.embeddings.utils import url_retrieve
+
+    image_url = "http://farm1.staticflickr.com/53/167798175_7c7845bbbd_z.jpg"
+    image_bytes = url_retrieve(image_url)
+    img = Image.open(io.BytesIO(image_bytes))
+    assert img.size[0] > 0 and img.size[1] > 0

python/python/tests/test_fts.py

@@ -27,6 +27,7 @@ from lancedb.query import (
     PhraseQuery,
     BooleanQuery,
     Occur,
+    LanceFtsQueryBuilder,
 )
 import numpy as np
 import pyarrow as pa
@@ -882,3 +883,109 @@ def test_fts_query_to_json():
         '"must_not":[]}}'
     )
     assert json_str == expected
+
+
+def test_fts_fast_search(table):
+    table.create_fts_index("text", use_tantivy=False)
+
+    # Insert some unindexed data
+    table.add(
+        [
+            {
+                "text": "xyz",
+                "vector": [0 for _ in range(128)],
+                "id": 101,
+                "text2": "xyz",
+                "nested": {"text": "xyz"},
+                "count": 10,
+            }
+        ]
+    )
+
+    # Without fast_search, the query object should not have fast_search set
+    builder = table.search("xyz", query_type="fts").limit(10)
+    query = builder.to_query_object()
+    assert query.fast_search is None
+
+    # With fast_search, the query object should have fast_search=True
+    builder = table.search("xyz", query_type="fts").fast_search().limit(10)
+    query = builder.to_query_object()
+    assert query.fast_search is True
+
+    # fast_search should be chainable with other methods
+    builder = (
+        table.search("xyz", query_type="fts").fast_search().select(["text"]).limit(5)
+    )
+    query = builder.to_query_object()
+    assert query.fast_search is True
+    assert query.limit == 5
+    assert query.columns == ["text"]
+
+    # fast_search should be enabled by keyword argument too
+    query = LanceFtsQueryBuilder(table, "xyz", fast_search=True).to_query_object()
+    assert query.fast_search is True
+
+    # Verify it executes without error and skips unindexed data
+    results = table.search("xyz", query_type="fts").fast_search().limit(5).to_list()
+    assert len(results) == 0
+
+    # Update index and verify it returns results
+    table.optimize()
+    results = table.search("xyz", query_type="fts").fast_search().limit(5).to_list()
+    assert len(results) > 0
+
+
+@pytest.mark.asyncio
+async def test_fts_fast_search_async(async_table):
+    await async_table.create_index("text", config=FTS())
+
+    # Insert some unindexed data
+    await async_table.add(
+        [
+            {
+                "text": "xyz",
+                "vector": [0 for _ in range(128)],
+                "id": 101,
+                "text2": "xyz",
+                "nested": {"text": "xyz"},
+                "count": 10,
+            }
+        ]
+    )
+
+    # Without fast_search, should return results
+    results = await async_table.query().nearest_to_text("xyz").limit(5).to_list()
+    assert len(results) > 0
+
+    # With fast_search, should return no results data unindexed
+    fast_results = (
+        await async_table.query()
+        .nearest_to_text("xyz")
+        .fast_search()
+        .limit(5)
+        .to_list()
+    )
+    assert len(fast_results) == 0
+
+    # Update index and verify it returns results
+    await async_table.optimize()
+
+    fast_results = (
+        await async_table.query()
+        .nearest_to_text("xyz")
+        .fast_search()
+        .limit(5)
+        .to_list()
+    )
+    assert len(fast_results) > 0
+
+    # fast_search should be chainable with other methods
+    results = (
+        await async_table.query()
+        .nearest_to_text("xyz")
+        .fast_search()
+        .select(["text"])
+        .limit(5)
+        .to_list()
+    )
+    assert len(results) > 0

python/python/tests/test_hybrid_query.py

@@ -177,6 +177,60 @@ async def test_analyze_plan(table: AsyncTable):
     assert "metrics=" in res
 
 
+@pytest.fixture
+def table_with_id(tmpdir_factory) -> Table:
+    tmp_path = str(tmpdir_factory.mktemp("data"))
+    db = lancedb.connect(tmp_path)
+    data = pa.table(
+        {
+            "id": pa.array([1, 2, 3, 4], type=pa.int64()),
+            "text": pa.array(["a", "b", "cat", "dog"]),
+            "vector": pa.array(
+                [[0.1, 0.1], [2, 2], [-0.1, -0.1], [0.5, -0.5]],
+                type=pa.list_(pa.float32(), list_size=2),
+            ),
+        }
+    )
+    table = db.create_table("test_with_id", data)
+    table.create_fts_index("text", with_position=False, use_tantivy=False)
+    return table
+
+
+def test_hybrid_prefilter_explain_plan(table_with_id: Table):
+    """
+    Verify that the prefilter logic is not inverted in LanceHybridQueryBuilder.
+    """
+    plan_prefilter = (
+        table_with_id.search(query_type="hybrid")
+        .vector([0.0, 0.0])
+        .text("dog")
+        .where("id = 1", prefilter=True)
+        .limit(2)
+        .explain_plan(verbose=True)
+    )
+
+    plan_postfilter = (
+        table_with_id.search(query_type="hybrid")
+        .vector([0.0, 0.0])
+        .text("dog")
+        .where("id = 1", prefilter=False)
+        .limit(2)
+        .explain_plan(verbose=True)
+    )
+
+    # prefilter=True: filter is pushed into the LanceRead scan.
+    # The FTS sub-plan exposes this as "full_filter=id = Int64(1)" inside LanceRead.
+    assert "full_filter=id = Int64(1)" in plan_prefilter, (
+        f"Should push the filter into the scan.\nPlan:\n{plan_prefilter}"
+    )
+
+    # prefilter=False: filter is applied as a separate FilterExec after the search.
+    # The filter must NOT be embedded in the scan.
+    assert "full_filter=id = Int64(1)" not in plan_postfilter, (
+        f"Should NOT push the filter into the scan.\nPlan:\n{plan_postfilter}"
+    )
+
+
 def test_normalize_scores():
     cases = [
         (pa.array([0.1, 0.4]), pa.array([0.0, 1.0])),

python/python/tests/test_index.py

@@ -3,6 +3,7 @@
 
 from datetime import timedelta
 import random
+from typing import get_args, get_type_hints
 
 import pyarrow as pa
 import pytest
@@ -22,6 +23,7 @@ from lancedb.index import (
     HnswSq,
     FTS,
 )
+from lancedb.table import IndexStatistics
 
 
 @pytest_asyncio.fixture
@@ -283,3 +285,23 @@ async def test_create_index_with_binary_vectors(binary_table: AsyncTable):
     for v in range(256):
         res = await binary_table.query().nearest_to([v] * 128).to_arrow()
         assert res["id"][0].as_py() == v
+
+
+def test_index_statistics_index_type_lists_all_supported_values():
+    expected_index_types = {
+        "IVF_FLAT",
+        "IVF_SQ",
+        "IVF_PQ",
+        "IVF_RQ",
+        "IVF_HNSW_SQ",
+        "IVF_HNSW_PQ",
+        "FTS",
+        "BTREE",
+        "BITMAP",
+        "LABEL_LIST",
+    }
+
+    assert (
+        set(get_args(get_type_hints(IndexStatistics)["index_type"]))
+        == expected_index_types
+    )

python/python/tests/test_namespace.py

@@ -8,6 +8,7 @@ import shutil
 import pytest
 import pyarrow as pa
 import lancedb
+from lance_namespace.errors import NamespaceNotEmptyError, TableNotFoundError
 
 
 class TestNamespaceConnection:
@@ -32,6 +33,16 @@ class TestNamespaceConnection:
         # Initially no tables in root
         assert len(list(db.table_names())) == 0
 
+    def test_connect_via_connect_helper(self):
+        """Connecting via lancedb.connect should delegate to namespace connection."""
+        db = lancedb.connect(
+            namespace_client_impl="dir",
+            namespace_client_properties={"root": self.temp_dir},
+        )
+
+        assert isinstance(db, lancedb.LanceNamespaceDBConnection)
+        assert len(list(db.table_names())) == 0
+
     def test_create_table_through_namespace(self):
         """Test creating a table through namespace."""
         db = lancedb.connect_namespace("dir", {"root": self.temp_dir})
@@ -49,14 +60,14 @@ class TestNamespaceConnection:
         )
 
         # Create empty table in child namespace
-        table = db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        table = db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
         assert table is not None
         assert table.name == "test_table"
         assert table.namespace == ["test_ns"]
         assert table.id == "test_ns$test_table"
 
         # Table should appear in child namespace
-        table_names = list(db.table_names(namespace=["test_ns"]))
+        table_names = list(db.table_names(namespace_path=["test_ns"]))
         assert "test_table" in table_names
         assert len(table_names) == 1
 
@@ -79,10 +90,10 @@ class TestNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
 
         # Open the table
-        table = db.open_table("test_table", namespace=["test_ns"])
+        table = db.open_table("test_table", namespace_path=["test_ns"])
         assert table is not None
         assert table.name == "test_table"
         assert table.namespace == ["test_ns"]
@@ -107,31 +118,31 @@ class TestNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        db.create_table("table1", schema=schema, namespace=["test_ns"])
+        db.create_table("table1", schema=schema, namespace_path=["test_ns"])
-        db.create_table("table2", schema=schema, namespace=["test_ns"])
+        db.create_table("table2", schema=schema, namespace_path=["test_ns"])
 
         # Verify both tables exist in child namespace
-        table_names = list(db.table_names(namespace=["test_ns"]))
+        table_names = list(db.table_names(namespace_path=["test_ns"]))
         assert "table1" in table_names
         assert "table2" in table_names
         assert len(table_names) == 2
 
         # Drop one table
-        db.drop_table("table1", namespace=["test_ns"])
+        db.drop_table("table1", namespace_path=["test_ns"])
 
         # Verify only table2 remains
-        table_names = list(db.table_names(namespace=["test_ns"]))
+        table_names = list(db.table_names(namespace_path=["test_ns"]))
         assert "table1" not in table_names
         assert "table2" in table_names
         assert len(table_names) == 1
 
         # Drop the second table
-        db.drop_table("table2", namespace=["test_ns"])
+        db.drop_table("table2", namespace_path=["test_ns"])
-        assert len(list(db.table_names(namespace=["test_ns"]))) == 0
+        assert len(list(db.table_names(namespace_path=["test_ns"]))) == 0
 
         # Should not be able to open dropped table
-        with pytest.raises(RuntimeError):
+        with pytest.raises(TableNotFoundError):
-            db.open_table("table1", namespace=["test_ns"])
+            db.open_table("table1", namespace_path=["test_ns"])
 
     def test_create_table_with_schema(self):
         """Test creating a table with explicit schema through namespace."""
@@ -150,7 +161,7 @@ class TestNamespaceConnection:
         )
 
         # Create table with schema in child namespace
-        table = db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        table = db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
         assert table is not None
         assert table.namespace == ["test_ns"]
 
@@ -174,7 +185,7 @@ class TestNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        db.create_table("old_name", schema=schema, namespace=["test_ns"])
+        db.create_table("old_name", schema=schema, namespace_path=["test_ns"])
 
         # Rename should raise NotImplementedError
         with pytest.raises(NotImplementedError, match="rename_table is not supported"):
@@ -195,20 +206,20 @@ class TestNamespaceConnection:
             ]
         )
         for i in range(3):
-            db.create_table(f"table{i}", schema=schema, namespace=["test_ns"])
+            db.create_table(f"table{i}", schema=schema, namespace_path=["test_ns"])
 
         # Verify tables exist in child namespace
-        assert len(list(db.table_names(namespace=["test_ns"]))) == 3
+        assert len(list(db.table_names(namespace_path=["test_ns"]))) == 3
 
         # Drop all tables in child namespace
-        db.drop_all_tables(namespace=["test_ns"])
+        db.drop_all_tables(namespace_path=["test_ns"])
 
         # Verify all tables are gone from child namespace
-        assert len(list(db.table_names(namespace=["test_ns"]))) == 0
+        assert len(list(db.table_names(namespace_path=["test_ns"]))) == 0
 
         # Test that table_names works with keyword-only namespace parameter
-        db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
-        result = list(db.table_names(namespace=["test_ns"]))
+        result = list(db.table_names(namespace_path=["test_ns"]))
         assert "test_table" in result
 
     def test_table_operations(self):
@@ -226,7 +237,7 @@ class TestNamespaceConnection:
                 pa.field("text", pa.string()),
             ]
         )
-        table = db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        table = db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
 
         # Verify empty table was created
         result = table.to_pandas()
@@ -297,25 +308,25 @@ class TestNamespaceConnection:
             ]
         )
         table = db.create_table(
-            "test_table", schema=schema, namespace=["test_namespace"]
+            "test_table", schema=schema, namespace_path=["test_namespace"]
         )
         assert table is not None
 
         # Verify table exists in namespace
-        tables_in_namespace = list(db.table_names(namespace=["test_namespace"]))
+        tables_in_namespace = list(db.table_names(namespace_path=["test_namespace"]))
         assert "test_table" in tables_in_namespace
         assert len(tables_in_namespace) == 1
 
         # Open table from namespace
-        table = db.open_table("test_table", namespace=["test_namespace"])
+        table = db.open_table("test_table", namespace_path=["test_namespace"])
         assert table is not None
         assert table.name == "test_table"
 
         # Drop table from namespace
-        db.drop_table("test_table", namespace=["test_namespace"])
+        db.drop_table("test_table", namespace_path=["test_namespace"])
 
         # Verify table no longer exists in namespace
-        tables_in_namespace = list(db.table_names(namespace=["test_namespace"]))
+        tables_in_namespace = list(db.table_names(namespace_path=["test_namespace"]))
         assert len(tables_in_namespace) == 0
 
         # Drop namespace
@@ -337,14 +348,14 @@ class TestNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        db.create_table("test_table", schema=schema, namespace=["test_namespace"])
+        db.create_table("test_table", schema=schema, namespace_path=["test_namespace"])
 
         # Try to drop namespace with tables - should fail
-        with pytest.raises(RuntimeError, match="is not empty"):
+        with pytest.raises(NamespaceNotEmptyError):
             db.drop_namespace(["test_namespace"])
 
         # Drop table first
-        db.drop_table("test_table", namespace=["test_namespace"])
+        db.drop_table("test_table", namespace_path=["test_namespace"])
 
         # Now dropping namespace should work
         db.drop_namespace(["test_namespace"])
@@ -367,10 +378,10 @@ class TestNamespaceConnection:
 
         # Create table with same name in both namespaces
         table_a = db.create_table(
-            "same_name_table", schema=schema, namespace=["namespace_a"]
+            "same_name_table", schema=schema, namespace_path=["namespace_a"]
         )
         table_b = db.create_table(
-            "same_name_table", schema=schema, namespace=["namespace_b"]
+            "same_name_table", schema=schema, namespace_path=["namespace_b"]
         )
 
         # Add different data to each table
@@ -388,7 +399,9 @@ class TestNamespaceConnection:
         table_b.add(data_b)
 
         # Verify data in namespace_a table
-        opened_table_a = db.open_table("same_name_table", namespace=["namespace_a"])
+        opened_table_a = db.open_table(
+            "same_name_table", namespace_path=["namespace_a"]
+        )
         result_a = opened_table_a.to_pandas().sort_values("id").reset_index(drop=True)
         assert len(result_a) == 2
         assert result_a["id"].tolist() == [1, 2]
@@ -399,7 +412,9 @@ class TestNamespaceConnection:
         assert [v.tolist() for v in result_a["vector"]] == [[1.0, 2.0], [3.0, 4.0]]
 
         # Verify data in namespace_b table
-        opened_table_b = db.open_table("same_name_table", namespace=["namespace_b"])
+        opened_table_b = db.open_table(
+            "same_name_table", namespace_path=["namespace_b"]
+        )
         result_b = opened_table_b.to_pandas().sort_values("id").reset_index(drop=True)
         assert len(result_b) == 3
         assert result_b["id"].tolist() == [10, 20, 30]
@@ -419,8 +434,8 @@ class TestNamespaceConnection:
         assert "same_name_table" not in root_tables
 
         # Clean up
-        db.drop_table("same_name_table", namespace=["namespace_a"])
+        db.drop_table("same_name_table", namespace_path=["namespace_a"])
-        db.drop_table("same_name_table", namespace=["namespace_b"])
+        db.drop_table("same_name_table", namespace_path=["namespace_b"])
         db.drop_namespace(["namespace_a"])
         db.drop_namespace(["namespace_b"])
 
@@ -448,6 +463,8 @@ class TestAsyncNamespaceConnection:
         table_names = await db.table_names()
         assert len(list(table_names)) == 0
 
+    # Async connect via namespace helper is not enabled yet.
+
     async def test_create_table_async(self):
         """Test creating a table asynchronously through namespace."""
         db = lancedb.connect_namespace_async("dir", {"root": self.temp_dir})
@@ -466,13 +483,13 @@ class TestAsyncNamespaceConnection:
 
         # Create empty table in child namespace
         table = await db.create_table(
-            "test_table", schema=schema, namespace=["test_ns"]
+            "test_table", schema=schema, namespace_path=["test_ns"]
         )
         assert table is not None
         assert isinstance(table, lancedb.AsyncTable)
 
         # Table should appear in child namespace
-        table_names = await db.table_names(namespace=["test_ns"])
+        table_names = await db.table_names(namespace_path=["test_ns"])
         assert "test_table" in list(table_names)
 
     async def test_open_table_async(self):
@@ -489,10 +506,10 @@ class TestAsyncNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        await db.create_table("test_table", schema=schema, namespace=["test_ns"])
+        await db.create_table("test_table", schema=schema, namespace_path=["test_ns"])
 
         # Open the table
-        table = await db.open_table("test_table", namespace=["test_ns"])
+        table = await db.open_table("test_table", namespace_path=["test_ns"])
         assert table is not None
         assert isinstance(table, lancedb.AsyncTable)
 
@@ -546,20 +563,20 @@ class TestAsyncNamespaceConnection:
                 pa.field("vector", pa.list_(pa.float32(), 2)),
             ]
         )
-        await db.create_table("table1", schema=schema, namespace=["test_ns"])
+        await db.create_table("table1", schema=schema, namespace_path=["test_ns"])
-        await db.create_table("table2", schema=schema, namespace=["test_ns"])
+        await db.create_table("table2", schema=schema, namespace_path=["test_ns"])
 
         # Verify both tables exist in child namespace
-        table_names = list(await db.table_names(namespace=["test_ns"]))
+        table_names = list(await db.table_names(namespace_path=["test_ns"]))
         assert "table1" in table_names
         assert "table2" in table_names
         assert len(table_names) == 2
 
         # Drop one table
-        await db.drop_table("table1", namespace=["test_ns"])
+        await db.drop_table("table1", namespace_path=["test_ns"])
 
         # Verify only table2 remains
-        table_names = list(await db.table_names(namespace=["test_ns"]))
+        table_names = list(await db.table_names(namespace_path=["test_ns"]))
         assert "table1" not in table_names
         assert "table2" in table_names
         assert len(table_names) == 1
@@ -588,20 +605,24 @@ class TestAsyncNamespaceConnection:
             ]
         )
         table = await db.create_table(
-            "test_table", schema=schema, namespace=["test_namespace"]
+            "test_table", schema=schema, namespace_path=["test_namespace"]
         )
         assert table is not None
 
         # Verify table exists in namespace
-        tables_in_namespace = list(await db.table_names(namespace=["test_namespace"]))
+        tables_in_namespace = list(
+            await db.table_names(namespace_path=["test_namespace"])
+        )
         assert "test_table" in tables_in_namespace
         assert len(tables_in_namespace) == 1
 
         # Drop table from namespace
-        await db.drop_table("test_table", namespace=["test_namespace"])
+        await db.drop_table("test_table", namespace_path=["test_namespace"])
 
         # Verify table no longer exists in namespace
-        tables_in_namespace = list(await db.table_names(namespace=["test_namespace"]))
+        tables_in_namespace = list(
+            await db.table_names(namespace_path=["test_namespace"])
+        )
         assert len(tables_in_namespace) == 0
 
         # Drop namespace
@@ -626,15 +647,98 @@ class TestAsyncNamespaceConnection:
             ]
         )
         for i in range(3):
-            await db.create_table(f"table{i}", schema=schema, namespace=["test_ns"])
+            await db.create_table(
+                f"table{i}", schema=schema, namespace_path=["test_ns"]
+            )
 
         # Verify tables exist in child namespace
-        table_names = await db.table_names(namespace=["test_ns"])
+        table_names = await db.table_names(namespace_path=["test_ns"])
         assert len(list(table_names)) == 3
 
         # Drop all tables in child namespace
-        await db.drop_all_tables(namespace=["test_ns"])
+        await db.drop_all_tables(namespace_path=["test_ns"])
 
         # Verify all tables are gone from child namespace
-        table_names = await db.table_names(namespace=["test_ns"])
+        table_names = await db.table_names(namespace_path=["test_ns"])
         assert len(list(table_names)) == 0
+
+
+class TestPushdownOperations:
+    """Test pushdown operations on namespace connections."""
+
+    def setup_method(self):
+        """Set up test fixtures."""
+        self.temp_dir = tempfile.mkdtemp()
+
+    def teardown_method(self):
+        """Clean up test fixtures."""
+        shutil.rmtree(self.temp_dir, ignore_errors=True)
+
+    def test_query_table_pushdown_stored(self):
+        """Test that QueryTable pushdown is stored on sync connection."""
+        db = lancedb.connect_namespace(
+            "dir",
+            {"root": self.temp_dir},
+            namespace_client_pushdown_operations=["QueryTable"],
+        )
+        assert "QueryTable" in db._pushdown_operations
+
+    def test_create_table_pushdown_stored(self):
+        """Test that CreateTable pushdown is stored on sync connection."""
+        db = lancedb.connect_namespace(
+            "dir",
+            {"root": self.temp_dir},
+            namespace_client_pushdown_operations=["CreateTable"],
+        )
+        assert "CreateTable" in db._pushdown_operations
+
+    def test_both_pushdowns_stored(self):
+        """Test that both pushdown operations can be set together."""
+        db = lancedb.connect_namespace(
+            "dir",
+            {"root": self.temp_dir},
+            namespace_client_pushdown_operations=["QueryTable", "CreateTable"],
+        )
+        assert "QueryTable" in db._pushdown_operations
+        assert "CreateTable" in db._pushdown_operations
+
+    def test_pushdown_defaults_to_empty(self):
+        """Test that pushdown operations default to empty."""
+        db = lancedb.connect_namespace("dir", {"root": self.temp_dir})
+        assert len(db._pushdown_operations) == 0
+
+
+@pytest.mark.asyncio
+class TestAsyncPushdownOperations:
+    """Test pushdown operations on async namespace connections."""
+
+    def setup_method(self):
+        """Set up test fixtures."""
+        self.temp_dir = tempfile.mkdtemp()
+
+    def teardown_method(self):
+        """Clean up test fixtures."""
+        shutil.rmtree(self.temp_dir, ignore_errors=True)
+
+    async def test_async_query_table_pushdown_stored(self):
+        """Test that QueryTable pushdown is stored on async connection."""
+        db = lancedb.connect_namespace_async(
+            "dir",
+            {"root": self.temp_dir},
+            namespace_client_pushdown_operations=["QueryTable"],
+        )
+        assert "QueryTable" in db._pushdown_operations
+
+    async def test_async_create_table_pushdown_stored(self):
+        """Test that CreateTable pushdown is stored on async connection."""
+        db = lancedb.connect_namespace_async(
+            "dir",
+            {"root": self.temp_dir},
+            namespace_client_pushdown_operations=["CreateTable"],
+        )
+        assert "CreateTable" in db._pushdown_operations
+
+    async def test_async_pushdown_defaults_to_empty(self):
+        """Test that pushdown operations default to empty on async connection."""
+        db = lancedb.connect_namespace_async("dir", {"root": self.temp_dir})
+        assert len(db._pushdown_operations) == 0

python/python/tests/test_namespace_integration.py

@@ -4,9 +4,11 @@
 """
 Integration tests for LanceDB Namespace with S3 and credential refresh.
 
-This test simulates a namespace server that returns incrementing credentials
+This test uses DirectoryNamespace with native ops_metrics and vend_input_storage_options
-and verifies that the credential refresh mechanism works correctly for both
+features to track API calls and test credential refresh mechanisms.
-create_table and open_table operations.
+
+Tests are parameterized to run with both DirectoryNamespace and a CustomNamespace
+wrapper to verify Python-Rust binding works correctly for custom implementations.
 
 Tests verify:
 - Storage options provider is auto-created and used
@@ -18,22 +20,136 @@ Tests verify:
 import copy
 import time
 import uuid
-from threading import Lock
+from typing import Dict, Optional
-from typing import Dict
 
 import pyarrow as pa
 import pytest
-from lance_namespace import (
+from lance.namespace import (
-    CreateEmptyTableRequest,
-    CreateEmptyTableResponse,
     DeclareTableRequest,
     DeclareTableResponse,
     DescribeTableRequest,
     DescribeTableResponse,
+    DirectoryNamespace,
     LanceNamespace,
 )
+from lance_namespace import (
+    CreateNamespaceRequest,
+    CreateNamespaceResponse,
+    CreateTableRequest,
+    CreateTableResponse,
+    CreateTableVersionRequest,
+    CreateTableVersionResponse,
+    DeregisterTableRequest,
+    DeregisterTableResponse,
+    DescribeNamespaceRequest,
+    DescribeNamespaceResponse,
+    DescribeTableVersionRequest,
+    DescribeTableVersionResponse,
+    DropNamespaceRequest,
+    DropNamespaceResponse,
+    DropTableRequest,
+    DropTableResponse,
+    ListNamespacesRequest,
+    ListNamespacesResponse,
+    ListTablesRequest,
+    ListTablesResponse,
+    ListTableVersionsRequest,
+    ListTableVersionsResponse,
+    NamespaceExistsRequest,
+    RegisterTableRequest,
+    RegisterTableResponse,
+    TableExistsRequest,
+)
 from lancedb.namespace import LanceNamespaceDBConnection
 
+
+class CustomNamespace(LanceNamespace):
+    """A custom namespace wrapper that delegates to DirectoryNamespace.
+
+    This class verifies that the Python-Rust binding works correctly for
+    custom namespace implementations that wrap the native DirectoryNamespace.
+    All methods simply delegate to the underlying DirectoryNamespace instance.
+    """
+
+    def __init__(self, inner: DirectoryNamespace):
+        self._inner = inner
+
+    def namespace_id(self) -> str:
+        return f"CustomNamespace[{self._inner.namespace_id()}]"
+
+    def create_namespace(
+        self, request: CreateNamespaceRequest
+    ) -> CreateNamespaceResponse:
+        return self._inner.create_namespace(request)
+
+    def describe_namespace(
+        self, request: DescribeNamespaceRequest
+    ) -> DescribeNamespaceResponse:
+        return self._inner.describe_namespace(request)
+
+    def namespace_exists(self, request: NamespaceExistsRequest) -> None:
+        return self._inner.namespace_exists(request)
+
+    def drop_namespace(self, request: DropNamespaceRequest) -> DropNamespaceResponse:
+        return self._inner.drop_namespace(request)
+
+    def list_namespaces(self, request: ListNamespacesRequest) -> ListNamespacesResponse:
+        return self._inner.list_namespaces(request)
+
+    def create_table(
+        self, request: CreateTableRequest, data: bytes
+    ) -> CreateTableResponse:
+        return self._inner.create_table(request, data)
+
+    def declare_table(self, request: DeclareTableRequest) -> DeclareTableResponse:
+        return self._inner.declare_table(request)
+
+    def describe_table(self, request: DescribeTableRequest) -> DescribeTableResponse:
+        return self._inner.describe_table(request)
+
+    def table_exists(self, request: TableExistsRequest) -> None:
+        return self._inner.table_exists(request)
+
+    def drop_table(self, request: DropTableRequest) -> DropTableResponse:
+        return self._inner.drop_table(request)
+
+    def list_tables(self, request: ListTablesRequest) -> ListTablesResponse:
+        return self._inner.list_tables(request)
+
+    def register_table(self, request: RegisterTableRequest) -> RegisterTableResponse:
+        return self._inner.register_table(request)
+
+    def deregister_table(
+        self, request: DeregisterTableRequest
+    ) -> DeregisterTableResponse:
+        return self._inner.deregister_table(request)
+
+    def list_table_versions(
+        self, request: ListTableVersionsRequest
+    ) -> ListTableVersionsResponse:
+        return self._inner.list_table_versions(request)
+
+    def describe_table_version(
+        self, request: DescribeTableVersionRequest
+    ) -> DescribeTableVersionResponse:
+        return self._inner.describe_table_version(request)
+
+    def create_table_version(
+        self, request: CreateTableVersionRequest
+    ) -> CreateTableVersionResponse:
+        return self._inner.create_table_version(request)
+
+    def retrieve_ops_metrics(self) -> Optional[Dict[str, int]]:
+        return self._inner.retrieve_ops_metrics()
+
+
+def _wrap_if_custom(ns_client: DirectoryNamespace, use_custom: bool):
+    """Wrap namespace client in CustomNamespace if use_custom is True."""
+    if use_custom:
+        return CustomNamespace(ns_client)
+    return ns_client
+
+
 # LocalStack S3 configuration
 CONFIG = {
     "allow_http": "true",
@@ -89,157 +205,88 @@ def delete_bucket(s3, bucket_name):
         pass
 
 
-class TrackingNamespace(LanceNamespace):
+def create_tracking_namespace(
+    bucket_name: str,
+    storage_options: dict,
+    credential_expires_in_seconds: int = 60,
+    use_custom: bool = False,
+):
+    """Create a DirectoryNamespace with ops metrics and credential vending enabled.
+
+    Uses native DirectoryNamespace features:
+    - ops_metrics_enabled=true: Tracks API call counts via retrieve_ops_metrics()
+    - vend_input_storage_options=true: Returns input storage options in responses
+    - vend_input_storage_options_refresh_interval_millis: Adds expires_at_millis
+
+    Args:
+        bucket_name: S3 bucket name or local path
+        storage_options: Storage options to pass through (credentials, endpoint, etc.)
+        credential_expires_in_seconds: Interval in seconds for credential expiration
+        use_custom: If True, wrap in CustomNamespace for testing custom implementations
+
+    Returns:
+        Tuple of (namespace_client, inner_namespace_client) where inner is always
+        the DirectoryNamespace (used for metrics retrieval)
     """
-    Mock namespace that wraps DirectoryNamespace and tracks API calls.
+    # Add refresh_offset_millis to storage options so that credentials are not
+    # considered expired immediately. Set to 1 second (1000ms) so that refresh
+    # checks work correctly with short-lived credentials in tests.
+    storage_options_with_refresh = dict(storage_options)
+    storage_options_with_refresh["refresh_offset_millis"] = "1000"
 
-    This namespace returns incrementing credentials with each API call to simulate
+    dir_props = {f"storage.{k}": v for k, v in storage_options_with_refresh.items()}
-    credential rotation. It also tracks the number of times each API is called
-    to verify caching behavior.
-    """
 
-    def __init__(
+    if bucket_name.startswith("/") or bucket_name.startswith("file://"):
-        self,
+        dir_props["root"] = f"{bucket_name}/namespace_root"
-        bucket_name: str,
+    else:
-        storage_options: Dict[str, str],
+        dir_props["root"] = f"s3://{bucket_name}/namespace_root"
-        credential_expires_in_seconds: int = 60,
-    ):
-        from lance.namespace import DirectoryNamespace
 
-        self.bucket_name = bucket_name
+    # Enable ops metrics tracking
-        self.base_storage_options = storage_options
+    dir_props["ops_metrics_enabled"] = "true"
-        self.credential_expires_in_seconds = credential_expires_in_seconds
+    # Enable storage options vending
-        self.describe_call_count = 0
+    dir_props["vend_input_storage_options"] = "true"
-        self.create_call_count = 0
+    # Set refresh interval in milliseconds
-        self.lock = Lock()
+    dir_props["vend_input_storage_options_refresh_interval_millis"] = str(
+        credential_expires_in_seconds * 1000
+    )
 
-        # Create underlying DirectoryNamespace with storage options
+    inner_ns_client = DirectoryNamespace(**dir_props)
-        dir_props = {f"storage.{k}": v for k, v in storage_options.items()}
+    ns_client = _wrap_if_custom(inner_ns_client, use_custom)
+    return ns_client, inner_ns_client
 
-        # Use S3 path for bucket name, local path for file paths
-        if bucket_name.startswith("/") or bucket_name.startswith("file://"):
-            dir_props["root"] = f"{bucket_name}/namespace_root"
-        else:
-            dir_props["root"] = f"s3://{bucket_name}/namespace_root"
 
-        self.inner = DirectoryNamespace(**dir_props)
+def get_describe_call_count(namespace_client) -> int:
+    """Get the number of describe_table calls made to the namespace client."""
+    return namespace_client.retrieve_ops_metrics().get("describe_table", 0)
 
-    def get_describe_call_count(self) -> int:
-        """Thread-safe getter for describe call count."""
-        with self.lock:
-            return self.describe_call_count
 
-    def get_create_call_count(self) -> int:
+def get_declare_call_count(namespace_client) -> int:
-        """Thread-safe getter for create call count."""
+    """Get the number of declare_table calls made to the namespace client."""
-        with self.lock:
+    return namespace_client.retrieve_ops_metrics().get("declare_table", 0)
-            return self.create_call_count
-
-    def namespace_id(self) -> str:
-        """Return namespace identifier."""
-        return f"TrackingNamespace {{ inner: {self.inner.namespace_id()} }}"
-
-    def _modify_storage_options(
-        self, storage_options: Dict[str, str], count: int
-    ) -> Dict[str, str]:
-        """
-        Add incrementing credentials with expiration timestamp.
-
-        This simulates a credential rotation system where each call returns
-        new credentials that expire after credential_expires_in_seconds.
-        """
-        modified = copy.deepcopy(storage_options) if storage_options else {}
-
-        # Increment credentials to simulate rotation
-        modified["aws_access_key_id"] = f"AKID_{count}"
-        modified["aws_secret_access_key"] = f"SECRET_{count}"
-        modified["aws_session_token"] = f"TOKEN_{count}"
-
-        # Set expiration time
-        expires_at_millis = int(
-            (time.time() + self.credential_expires_in_seconds) * 1000
-        )
-        modified["expires_at_millis"] = str(expires_at_millis)
-
-        return modified
-
-    def declare_table(self, request: DeclareTableRequest) -> DeclareTableResponse:
-        """Track declare_table calls and inject rotating credentials."""
-        with self.lock:
-            self.create_call_count += 1
-            count = self.create_call_count
-
-        response = self.inner.declare_table(request)
-        response.storage_options = self._modify_storage_options(
-            response.storage_options, count
-        )
-
-        return response
-
-    def create_empty_table(
-        self, request: CreateEmptyTableRequest
-    ) -> CreateEmptyTableResponse:
-        """Track create_empty_table calls and inject rotating credentials."""
-        with self.lock:
-            self.create_call_count += 1
-            count = self.create_call_count
-
-        response = self.inner.create_empty_table(request)
-        response.storage_options = self._modify_storage_options(
-            response.storage_options, count
-        )
-
-        return response
-
-    def describe_table(self, request: DescribeTableRequest) -> DescribeTableResponse:
-        """Track describe_table calls and inject rotating credentials."""
-        with self.lock:
-            self.describe_call_count += 1
-            count = self.describe_call_count
-
-        response = self.inner.describe_table(request)
-        response.storage_options = self._modify_storage_options(
-            response.storage_options, count
-        )
-
-        return response
-
-    # Pass through other methods to inner namespace
-    def list_tables(self, request):
-        return self.inner.list_tables(request)
-
-    def drop_table(self, request):
-        return self.inner.drop_table(request)
-
-    def list_namespaces(self, request):
-        return self.inner.list_namespaces(request)
-
-    def create_namespace(self, request):
-        return self.inner.create_namespace(request)
-
-    def drop_namespace(self, request):
-        return self.inner.drop_namespace(request)
 
 
 @pytest.mark.s3_test
-def test_namespace_create_table_with_provider(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_create_table_with_provider(s3_bucket: str, use_custom: bool):
     """
     Test creating a table through namespace with storage options provider.
 
     Verifies:
-    - create_empty_table is called once to reserve location
+    - declare_table is called once to reserve location
     - Storage options provider is auto-created
     - Table can be written successfully
     - Credentials are cached during write operations
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3600,  # 1 hour
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -249,8 +296,8 @@ def test_namespace_create_table_with_provider(s3_bucket: str):
     namespace_path = [namespace_name]
 
     # Verify initial state
-    assert namespace.get_create_call_count() == 0
+    assert get_declare_call_count(inner_ns_client) == 0
-    assert namespace.get_describe_call_count() == 0
+    assert get_describe_call_count(inner_ns_client) == 0
 
     # Create table with data
     data = pa.table(
@@ -261,12 +308,12 @@ def test_namespace_create_table_with_provider(s3_bucket: str):
         }
     )
 
-    table = db.create_table(table_name, data, namespace=namespace_path)
+    table = db.create_table(table_name, data, namespace_path=namespace_path)
 
-    # Verify create_empty_table was called exactly once
+    # Verify declare_table was called exactly once
-    assert namespace.get_create_call_count() == 1
+    assert get_declare_call_count(inner_ns_client) == 1
     # describe_table should NOT be called during create in create mode
-    assert namespace.get_describe_call_count() == 0
+    assert get_describe_call_count(inner_ns_client) == 0
 
     # Verify table was created successfully
     assert table.name == table_name
@@ -276,7 +323,8 @@ def test_namespace_create_table_with_provider(s3_bucket: str):
 
 
 @pytest.mark.s3_test
-def test_namespace_open_table_with_provider(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_open_table_with_provider(s3_bucket: str, use_custom: bool):
     """
     Test opening a table through namespace with storage options provider.
 
@@ -288,13 +336,14 @@ def test_namespace_open_table_with_provider(s3_bucket: str):
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3600,
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -312,21 +361,21 @@ def test_namespace_open_table_with_provider(s3_bucket: str):
         }
     )
 
-    db.create_table(table_name, data, namespace=namespace_path)
+    db.create_table(table_name, data, namespace_path=namespace_path)
 
-    initial_create_count = namespace.get_create_call_count()
+    initial_declare_count = get_declare_call_count(inner_ns_client)
-    assert initial_create_count == 1
+    assert initial_declare_count == 1
 
     # Open the table
-    opened_table = db.open_table(table_name, namespace=namespace_path)
+    opened_table = db.open_table(table_name, namespace_path=namespace_path)
 
     # Verify describe_table was called exactly once
-    assert namespace.get_describe_call_count() == 1
+    assert get_describe_call_count(inner_ns_client) == 1
-    # create_empty_table should not be called again
+    # declare_table should not be called again
-    assert namespace.get_create_call_count() == initial_create_count
+    assert get_declare_call_count(inner_ns_client) == initial_declare_count
 
     # Perform multiple read operations
-    describe_count_after_open = namespace.get_describe_call_count()
+    describe_count_after_open = get_describe_call_count(inner_ns_client)
 
     for _ in range(3):
         result = opened_table.to_pandas()
@@ -335,11 +384,12 @@ def test_namespace_open_table_with_provider(s3_bucket: str):
         assert count == 5
 
     # Verify credentials were cached (no additional describe_table calls)
-    assert namespace.get_describe_call_count() == describe_count_after_open
+    assert get_describe_call_count(inner_ns_client) == describe_count_after_open
 
 
 @pytest.mark.s3_test
-def test_namespace_credential_refresh_on_read(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_credential_refresh_on_read(s3_bucket: str, use_custom: bool):
     """
     Test credential refresh when credentials expire during read operations.
 
@@ -350,13 +400,14 @@ def test_namespace_credential_refresh_on_read(s3_bucket: str):
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3,  # Short expiration for testing
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -373,16 +424,16 @@ def test_namespace_credential_refresh_on_read(s3_bucket: str):
         }
     )
 
-    db.create_table(table_name, data, namespace=namespace_path)
+    db.create_table(table_name, data, namespace_path=namespace_path)
 
     # Open table (triggers describe_table)
-    opened_table = db.open_table(table_name, namespace=namespace_path)
+    opened_table = db.open_table(table_name, namespace_path=namespace_path)
 
     # Perform an immediate read (should use credentials from open)
     result = opened_table.to_pandas()
     assert len(result) == 3
 
-    describe_count_after_first_read = namespace.get_describe_call_count()
+    describe_count_after_first_read = get_describe_call_count(inner_ns_client)
 
     # Wait for credentials to expire (3 seconds + buffer)
     time.sleep(5)
@@ -391,7 +442,7 @@ def test_namespace_credential_refresh_on_read(s3_bucket: str):
     result = opened_table.to_pandas()
     assert len(result) == 3
 
-    describe_count_after_refresh = namespace.get_describe_call_count()
+    describe_count_after_refresh = get_describe_call_count(inner_ns_client)
     # Verify describe_table was called again (credential refresh)
     refresh_delta = describe_count_after_refresh - describe_count_after_first_read
 
@@ -404,7 +455,8 @@ def test_namespace_credential_refresh_on_read(s3_bucket: str):
 
 
 @pytest.mark.s3_test
-def test_namespace_credential_refresh_on_write(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_credential_refresh_on_write(s3_bucket: str, use_custom: bool):
     """
     Test credential refresh when credentials expire during write operations.
 
@@ -415,13 +467,14 @@ def test_namespace_credential_refresh_on_write(s3_bucket: str):
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3,  # Short expiration
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -438,7 +491,7 @@ def test_namespace_credential_refresh_on_write(s3_bucket: str):
         }
     )
 
-    table = db.create_table(table_name, initial_data, namespace=namespace_path)
+    table = db.create_table(table_name, initial_data, namespace_path=namespace_path)
 
     # Add more data (should use cached credentials)
     new_data = pa.table(
@@ -466,24 +519,26 @@ def test_namespace_credential_refresh_on_write(s3_bucket: str):
 
 
 @pytest.mark.s3_test
-def test_namespace_overwrite_mode(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_overwrite_mode(s3_bucket: str, use_custom: bool):
     """
     Test creating table in overwrite mode with credential tracking.
 
     Verifies:
-    - First create calls create_empty_table exactly once
+    - First create calls declare_table exactly once
     - Overwrite mode calls describe_table exactly once to check existence
     - Storage options provider works in overwrite mode
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3600,
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -500,11 +555,11 @@ def test_namespace_overwrite_mode(s3_bucket: str):
         }
     )
 
-    table = db.create_table(table_name, data1, namespace=namespace_path)
+    table = db.create_table(table_name, data1, namespace_path=namespace_path)
-    # Exactly one create_empty_table call for initial create
+    # Exactly one declare_table call for initial create
-    assert namespace.get_create_call_count() == 1
+    assert get_declare_call_count(inner_ns_client) == 1
     # No describe_table calls in create mode
-    assert namespace.get_describe_call_count() == 0
+    assert get_describe_call_count(inner_ns_client) == 0
     assert table.count_rows() == 2
 
     # Overwrite the table
@@ -516,14 +571,14 @@ def test_namespace_overwrite_mode(s3_bucket: str):
     )
 
     table2 = db.create_table(
-        table_name, data2, namespace=namespace_path, mode="overwrite"
+        table_name, data2, namespace_path=namespace_path, mode="overwrite"
     )
 
-    # Should still have only 1 create_empty_table call
+    # Should still have only 1 declare_table call
     # (overwrite reuses location from describe_table)
-    assert namespace.get_create_call_count() == 1
+    assert get_declare_call_count(inner_ns_client) == 1
     # Should have called describe_table exactly once to get existing table location
-    assert namespace.get_describe_call_count() == 1
+    assert get_describe_call_count(inner_ns_client) == 1
 
     # Verify new data
     assert table2.count_rows() == 3
@@ -532,7 +587,8 @@ def test_namespace_overwrite_mode(s3_bucket: str):
 
 
 @pytest.mark.s3_test
-def test_namespace_multiple_tables(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_multiple_tables(s3_bucket: str, use_custom: bool):
     """
     Test creating and opening multiple tables in the same namespace.
 
@@ -543,13 +599,14 @@ def test_namespace_multiple_tables(s3_bucket: str):
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3600,
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -559,22 +616,22 @@ def test_namespace_multiple_tables(s3_bucket: str):
     # Create first table
     table1_name = f"table1_{uuid.uuid4().hex}"
     data1 = pa.table({"id": [1, 2], "value": [10, 20]})
-    db.create_table(table1_name, data1, namespace=namespace_path)
+    db.create_table(table1_name, data1, namespace_path=namespace_path)
 
     # Create second table
     table2_name = f"table2_{uuid.uuid4().hex}"
     data2 = pa.table({"id": [3, 4], "value": [30, 40]})
-    db.create_table(table2_name, data2, namespace=namespace_path)
+    db.create_table(table2_name, data2, namespace_path=namespace_path)
 
-    # Should have 2 create calls (one per table)
+    # Should have 2 declare calls (one per table)
-    assert namespace.get_create_call_count() == 2
+    assert get_declare_call_count(inner_ns_client) == 2
 
     # Open both tables
-    opened1 = db.open_table(table1_name, namespace=namespace_path)
+    opened1 = db.open_table(table1_name, namespace_path=namespace_path)
-    opened2 = db.open_table(table2_name, namespace=namespace_path)
+    opened2 = db.open_table(table2_name, namespace_path=namespace_path)
 
     # Should have 2 describe calls (one per open)
-    assert namespace.get_describe_call_count() == 2
+    assert get_describe_call_count(inner_ns_client) == 2
 
     # Verify both tables work independently
     assert opened1.count_rows() == 2
@@ -588,7 +645,8 @@ def test_namespace_multiple_tables(s3_bucket: str):
 
 
 @pytest.mark.s3_test
-def test_namespace_with_schema_only(s3_bucket: str):
+@pytest.mark.parametrize("use_custom", [False, True], ids=["DirectoryNS", "CustomNS"])
+def test_namespace_with_schema_only(s3_bucket: str, use_custom: bool):
     """
     Test creating empty table with schema only (no data).
 
@@ -599,13 +657,14 @@ def test_namespace_with_schema_only(s3_bucket: str):
     """
     storage_options = copy.deepcopy(CONFIG)
 
-    namespace = TrackingNamespace(
+    ns_client, inner_ns_client = create_tracking_namespace(
         bucket_name=s3_bucket,
         storage_options=storage_options,
         credential_expires_in_seconds=3600,
+        use_custom=use_custom,
     )
 
-    db = LanceNamespaceDBConnection(namespace)
+    db = LanceNamespaceDBConnection(ns_client)
 
     # Create unique namespace for this test
     namespace_name = f"test_ns_{uuid.uuid4().hex[:8]}"
@@ -623,12 +682,12 @@ def test_namespace_with_schema_only(s3_bucket: str):
         ]
     )
 
-    table = db.create_table(table_name, schema=schema, namespace=namespace_path)
+    table = db.create_table(table_name, schema=schema, namespace_path=namespace_path)
 
-    # Should have called create_empty_table once
+    # Should have called declare_table once
-    assert namespace.get_create_call_count() == 1
+    assert get_declare_call_count(inner_ns_client) == 1
     # Should NOT have called describe_table in create mode
-    assert namespace.get_describe_call_count() == 0
+    assert get_describe_call_count(inner_ns_client) == 0
 
     # Verify empty table
     assert table.count_rows() == 0

python/python/tests/test_query.py

@@ -30,6 +30,7 @@ from lancedb.query import (
     PhraseQuery,
     Query,
     FullTextSearchQuery,
+    ensure_vector_query,
 )
 from lancedb.rerankers.cross_encoder import CrossEncoderReranker
 from lancedb.table import AsyncTable, LanceTable
@@ -1501,6 +1502,18 @@ def test_search_empty_table(mem_db):
     assert results == []
 
 
+def test_ensure_vector_query_empty_list():
+    """Regression: ensure_vector_query used to return instead of raise ValueError."""
+    with pytest.raises(ValueError, match="non-empty"):
+        ensure_vector_query([])
+
+
+def test_ensure_vector_query_nested_empty_list():
+    """Regression: ensure_vector_query used to return instead of raise ValueError."""
+    with pytest.raises(ValueError, match="non-empty"):
+        ensure_vector_query([[]])
+
+
 def test_fast_search(tmp_path):
     db = lancedb.connect(tmp_path)
 

python/python/tests/test_remote_db.py

@@ -1201,6 +1201,18 @@ async def test_header_provider_overrides_static_headers():
         await db.table_names()
 
 
+def test_close():
+    """Test that close() works without AttributeError."""
+    import asyncio
+
+    def handler(req):
+        req.send_response(200)
+        req.end_headers()
+
+    with mock_lancedb_connection(handler) as db:
+        asyncio.run(db.close())
+
+
 @pytest.mark.parametrize("exception", [KeyboardInterrupt, SystemExit, GeneratorExit])
 def test_background_loop_cancellation(exception):
     """Test that BackgroundEventLoop.run() cancels the future on interrupt."""

python/python/tests/test_table.py

@@ -326,6 +326,24 @@ def test_add_struct(mem_db: DBConnection):
     table = mem_db.create_table("test2", schema=schema)
     table.add(data)
 
+    struct_type = pa.struct(
+        [
+            ("b", pa.int64()),
+            ("a", pa.int64()),
+        ]
+    )
+    expected = pa.table(
+        {
+            "s_list": [
+                [
+                    pa.scalar({"b": 1, "a": 2}, type=struct_type),
+                    pa.scalar({"b": 4, "a": None}, type=struct_type),
+                ]
+            ],
+        }
+    )
+    assert table.to_arrow() == expected
+
 
 def test_add_subschema(mem_db: DBConnection):
     schema = pa.schema(
@@ -509,6 +527,132 @@ async def test_add_async(mem_db_async: AsyncConnection):
     assert await table.count_rows() == 3
 
 
+def test_add_overwrite_infers_vector_schema(mem_db: DBConnection):
+    """Overwrite should infer vector columns the same way create_table does.
+
+    Regression test for https://github.com/lancedb/lancedb/issues/3183
+    """
+    table = mem_db.create_table(
+        "test_overwrite_vec",
+        data=[
+            {"vector": [1.0, 2.0, 3.0, 4.0], "item": "foo"},
+            {"vector": [5.0, 6.0, 7.0, 8.0], "item": "bar"},
+        ],
+    )
+    # create_table infers vector as fixed_size_list<float32, 4>
+    original_type = table.schema.field("vector").type
+    assert pa.types.is_fixed_size_list(original_type)
+
+    # overwrite with plain Python lists (PyArrow infers list<double>)
+    table.add(
+        [
+            {"vector": [10.0, 20.0, 30.0, 40.0], "item": "baz"},
+        ],
+        mode="overwrite",
+    )
+    # overwrite should infer vector column the same way as create_table
+    new_type = table.schema.field("vector").type
+    assert pa.types.is_fixed_size_list(new_type), (
+        f"Expected fixed_size_list after overwrite, got {new_type}"
+    )
+
+
+def test_add_progress_callback(mem_db: DBConnection):
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    updates = []
+    table.add([{"id": 3}, {"id": 4}], progress=lambda p: updates.append(dict(p)))
+
+    assert len(table) == 4
+    # The done callback always fires, so we should always get at least one.
+    assert len(updates) >= 1, "expected at least one progress callback"
+    for p in updates:
+        assert "output_rows" in p
+        assert "output_bytes" in p
+        assert "total_rows" in p
+        assert "elapsed_seconds" in p
+        assert "active_tasks" in p
+        assert "total_tasks" in p
+        assert "done" in p
+    # The last callback should have done=True.
+    assert updates[-1]["done"] is True
+
+
+def test_add_progress_tqdm_like(mem_db: DBConnection):
+    """Test that a tqdm-like object gets total set and update() called."""
+
+    class FakeBar:
+        def __init__(self):
+            self.total = None
+            self.n = 0
+            self.postfix = None
+
+        def update(self, n):
+            self.n += n
+
+        def set_postfix_str(self, s):
+            self.postfix = s
+
+        def refresh(self):
+            pass
+
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    bar = FakeBar()
+    table.add([{"id": 3}, {"id": 4}], progress=bar)
+
+    assert len(table) == 4
+    # Postfix should contain throughput and worker count
+    if bar.postfix is not None:
+        assert "MB/s" in bar.postfix
+        assert "workers" in bar.postfix
+
+
+def test_add_progress_bool(mem_db: DBConnection):
+    """Test that progress=True creates and closes a tqdm bar automatically."""
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    table.add([{"id": 3}, {"id": 4}], progress=True)
+    assert len(table) == 4
+
+    # progress=False should be the same as None
+    table.add([{"id": 5}], progress=False)
+    assert len(table) == 5
+
+
+@pytest.mark.asyncio
+async def test_add_progress_callback_async(mem_db_async: AsyncConnection):
+    """Progress callbacks work through the async path too."""
+    table = await mem_db_async.create_table("test", data=[{"id": 1}, {"id": 2}])
+
+    updates = []
+    await table.add([{"id": 3}, {"id": 4}], progress=lambda p: updates.append(dict(p)))
+
+    assert await table.count_rows() == 4
+    assert len(updates) >= 1
+    assert updates[-1]["done"] is True
+
+
+def test_add_progress_callback_error(mem_db: DBConnection):
+    """A failing callback must not prevent the write from succeeding."""
+    table = mem_db.create_table("test", data=[{"id": 1}, {"id": 2}])
+
+    def bad_callback(p):
+        raise RuntimeError("boom")
+
+    table.add([{"id": 3}, {"id": 4}], progress=bad_callback)
+    assert len(table) == 4
+
+
 def test_polars(mem_db: DBConnection):
     data = {
         "vector": [[3.1, 4.1], [5.9, 26.5]],
@@ -810,7 +954,7 @@ def test_create_index_name_and_train_parameters(
     )
 
 
-def test_add_with_nans(mem_db: DBConnection):
+def test_create_with_nans(mem_db: DBConnection):
     # by default we raise an error on bad input vectors
     bad_data = [
         {"vector": [np.nan], "item": "bar", "price": 20.0},
@@ -854,6 +998,57 @@ def test_add_with_nans(mem_db: DBConnection):
     assert np.allclose(v, np.array([0.0, 0.0]))
 
 
+def test_add_with_nans(mem_db: DBConnection):
+    schema = pa.schema(
+        [
+            pa.field("vector", pa.list_(pa.float32(), 2), nullable=True),
+            pa.field("item", pa.string(), nullable=True),
+            pa.field("price", pa.float64(), nullable=False),
+        ],
+    )
+    table = mem_db.create_table("test", schema=schema)
+    # by default we raise an error on bad input vectors
+    bad_data = [
+        {"vector": [np.nan], "item": "bar", "price": 20.0},
+        {"vector": [5], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, 5.0], "item": "bar", "price": 20.0},
+    ]
+    for row in bad_data:
+        with pytest.raises(ValueError):
+            table.add(
+                data=[row],
+            )
+
+    table.add(
+        [
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [2.1, 4.1], "item": "foo", "price": 9.0},
+            {"vector": [np.nan], "item": "bar", "price": 20.0},
+            {"vector": [5], "item": "bar", "price": 20.0},
+            {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        ],
+        on_bad_vectors="drop",
+    )
+    assert len(table) == 2
+    table.delete("true")
+
+    # We can fill bad input with some value
+    table.add(
+        data=[
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [np.nan], "item": "bar", "price": 20.0},
+            {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        ],
+        on_bad_vectors="fill",
+        fill_value=0.0,
+    )
+    assert len(table) == 3
+    arrow_tbl = table.search().where("item == 'bar'").to_arrow()
+    v = arrow_tbl["vector"].to_pylist()[0]
+    assert np.allclose(v, np.array([0.0, 0.0]))
+
+
 def test_restore(mem_db: DBConnection):
     table = mem_db.create_table(
         "my_table",
@@ -1913,7 +2108,7 @@ def test_stats(mem_db: DBConnection):
     stats = table.stats()
     print(f"{stats=}")
     assert stats == {
-        "total_bytes": 38,
+        "total_bytes": 60,
         "num_rows": 2,
         "num_indices": 0,
         "fragment_stats": {
@@ -1978,3 +2173,33 @@ def test_table_uri(tmp_path):
     db = lancedb.connect(tmp_path)
     table = db.create_table("my_table", data=[{"x": 0}])
     assert table.uri == str(tmp_path / "my_table.lance")
+
+
+def test_sanitize_data_metadata_not_stripped():
+    """Regression test: dict.update() returns None, so assigning its result
+    would silently replace metadata with None, causing with_metadata(None)
+    to strip all schema metadata from the target schema."""
+    from lancedb.table import _sanitize_data
+
+    schema = pa.schema(
+        [pa.field("x", pa.int64())],
+        metadata={b"existing_key": b"existing_value"},
+    )
+    batch = pa.record_batch([pa.array([1, 2, 3])], schema=schema)
+
+    # Use a different field type so the reader and target schemas differ,
+    # forcing _cast_to_target_schema to rebuild the schema with the
+    # target's metadata (instead of taking the fast-path).
+    target_schema = pa.schema(
+        [pa.field("x", pa.int32())],
+        metadata={b"existing_key": b"existing_value"},
+    )
+
+    reader = pa.RecordBatchReader.from_batches(schema, [batch])
+    metadata = {b"new_key": b"new_value"}
+    result = _sanitize_data(reader, target_schema=target_schema, metadata=metadata)
+
+    result_schema = result.schema
+    assert result_schema.metadata is not None
+    assert result_schema.metadata[b"existing_key"] == b"existing_value"
+    assert result_schema.metadata[b"new_key"] == b"new_value"

python/python/tests/test_util.py

@@ -121,6 +121,32 @@ def test_value_to_sql_string(tmp_path):
         assert table.to_pandas().query("search == @value")["replace"].item() == value
 
 
+def test_value_to_sql_dict():
+    # Simple flat struct
+    assert value_to_sql({"a": 1, "b": "hello"}) == "named_struct('a', 1, 'b', 'hello')"
+
+    # Nested struct
+    assert (
+        value_to_sql({"outer": {"inner": 1}})
+        == "named_struct('outer', named_struct('inner', 1))"
+    )
+
+    # List inside struct
+    assert value_to_sql({"a": [1, 2]}) == "named_struct('a', [1, 2])"
+
+    # Mixed types
+    assert (
+        value_to_sql({"name": "test", "count": 42, "rate": 3.14, "active": True})
+        == "named_struct('name', 'test', 'count', 42, 'rate', 3.14, 'active', TRUE)"
+    )
+
+    # Null value inside struct
+    assert value_to_sql({"a": None}) == "named_struct('a', NULL)"
+
+    # Empty dict
+    assert value_to_sql({}) == "named_struct()"
+
+
 def test_append_vector_columns():
     registry = EmbeddingFunctionRegistry.get_instance()
     registry.register("test")(MockTextEmbeddingFunction)

python/src/arrow.rs

@@ -10,7 +10,7 @@ use arrow::{
 use futures::stream::StreamExt;
 use lancedb::arrow::SendableRecordBatchStream;
 use pyo3::{
-    exceptions::PyStopAsyncIteration, pyclass, pymethods, Bound, Py, PyAny, PyRef, PyResult, Python,
+    Bound, Py, PyAny, PyRef, PyResult, Python, exceptions::PyStopAsyncIteration, pyclass, pymethods,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 

python/src/connection.rs

@@ -9,15 +9,17 @@ use lancedb::{
     database::{CreateTableMode, Database, ReadConsistency},
 };
 use pyo3::{
+    Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
     exceptions::{PyRuntimeError, PyValueError},
     pyclass, pyfunction, pymethods,
     types::{PyDict, PyDictMethods},
-    Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 
 use crate::{
-    error::PythonErrorExt, storage_options::py_object_to_storage_options_provider, table::Table,
+    error::PythonErrorExt,
+    namespace::{create_namespace_storage_options_provider, extract_namespace_arc},
+    table::Table,
 };
 
 #[pyclass]
@@ -86,16 +88,16 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (namespace=vec![], start_after=None, limit=None))]
+    #[pyo3(signature = (namespace_path=None, start_after=None, limit=None))]
     pub fn table_names(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         start_after: Option<String>,
         limit: Option<u32>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
         let mut op = inner.table_names();
-        op = op.namespace(namespace);
+        op = op.namespace(namespace_path.unwrap_or_default());
         if let Some(start_after) = start_after {
             op = op.start_after(start_after);
         }
@@ -106,34 +108,43 @@ impl Connection {
     }
 
     #[allow(clippy::too_many_arguments)]
-    #[pyo3(signature = (name, mode, data, namespace=vec![], storage_options=None, storage_options_provider=None, location=None))]
+    #[pyo3(signature = (name, mode, data, namespace_path=None, storage_options=None, location=None, namespace_client=None))]
     pub fn create_table<'a>(
         self_: PyRef<'a, Self>,
         name: String,
         mode: &str,
         data: Bound<'_, PyAny>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
-        storage_options_provider: Option<Py<PyAny>>,
         location: Option<String>,
+        namespace_client: Option<Py<PyAny>>,
     ) -> PyResult<Bound<'a, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let py = self_.py();
 
         let mode = Self::parse_create_mode_str(mode)?;
 
         let batches: Box<dyn arrow::array::RecordBatchReader + Send> =
             Box::new(ArrowArrayStreamReader::from_pyarrow_bound(&data)?);
 
-        let mut builder = inner.create_table(name, batches).mode(mode);
+        let ns_path = namespace_path.clone().unwrap_or_default();
+        let mut builder = inner.create_table(name.clone(), batches).mode(mode);
 
-        builder = builder.namespace(namespace);
+        builder = builder.namespace(ns_path.clone());
         if let Some(storage_options) = storage_options {
             builder = builder.storage_options(storage_options);
         }
-        if let Some(provider_obj) = storage_options_provider {
+
-            let provider = py_object_to_storage_options_provider(provider_obj)?;
+        // Auto-create storage options provider from namespace_client
+        if let Some(ns_obj) = namespace_client {
+            let ns_client = extract_namespace_arc(py, ns_obj)?;
+            // Create table_id by combining namespace_path with table name
+            let mut table_id = ns_path;
+            table_id.push(name);
+            let provider = create_namespace_storage_options_provider(ns_client, table_id);
             builder = builder.storage_options_provider(provider);
         }
+
         if let Some(location) = location {
             builder = builder.location(location);
         }
@@ -145,33 +156,44 @@ impl Connection {
     }
 
     #[allow(clippy::too_many_arguments)]
-    #[pyo3(signature = (name, mode, schema, namespace=vec![], storage_options=None, storage_options_provider=None, location=None))]
+    #[pyo3(signature = (name, mode, schema, namespace_path=None, storage_options=None, location=None, namespace_client=None))]
     pub fn create_empty_table<'a>(
         self_: PyRef<'a, Self>,
         name: String,
         mode: &str,
         schema: Bound<'_, PyAny>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
-        storage_options_provider: Option<Py<PyAny>>,
         location: Option<String>,
+        namespace_client: Option<Py<PyAny>>,
     ) -> PyResult<Bound<'a, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let py = self_.py();
 
         let mode = Self::parse_create_mode_str(mode)?;
 
         let schema = Schema::from_pyarrow_bound(&schema)?;
 
-        let mut builder = inner.create_empty_table(name, Arc::new(schema)).mode(mode);
+        let ns_path = namespace_path.clone().unwrap_or_default();
+        let mut builder = inner
+            .create_empty_table(name.clone(), Arc::new(schema))
+            .mode(mode);
 
-        builder = builder.namespace(namespace);
+        builder = builder.namespace(ns_path.clone());
         if let Some(storage_options) = storage_options {
             builder = builder.storage_options(storage_options);
         }
-        if let Some(provider_obj) = storage_options_provider {
+
-            let provider = py_object_to_storage_options_provider(provider_obj)?;
+        // Auto-create storage options provider from namespace_client
+        if let Some(ns_obj) = namespace_client {
+            let ns_client = extract_namespace_arc(py, ns_obj)?;
+            // Create table_id by combining namespace_path with table name
+            let mut table_id = ns_path;
+            table_id.push(name);
+            let provider = create_namespace_storage_options_provider(ns_client, table_id);
             builder = builder.storage_options_provider(provider);
         }
+
         if let Some(location) = location {
             builder = builder.location(location);
         }
@@ -182,33 +204,49 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (name, namespace=vec![], storage_options = None, storage_options_provider=None, index_cache_size = None, location=None))]
+    #[allow(clippy::too_many_arguments)]
+    #[pyo3(signature = (name, namespace_path=None, storage_options=None, index_cache_size=None, location=None, namespace_client=None, managed_versioning=None))]
     pub fn open_table(
         self_: PyRef<'_, Self>,
         name: String,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         storage_options: Option<HashMap<String, String>>,
-        storage_options_provider: Option<Py<PyAny>>,
         index_cache_size: Option<u32>,
         location: Option<String>,
+        namespace_client: Option<Py<PyAny>>,
+        managed_versioning: Option<bool>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let py = self_.py();
 
-        let mut builder = inner.open_table(name);
+        let ns_path = namespace_path.clone().unwrap_or_default();
-        builder = builder.namespace(namespace);
+        let mut builder = inner.open_table(name.clone());
+        builder = builder.namespace(ns_path.clone());
         if let Some(storage_options) = storage_options {
             builder = builder.storage_options(storage_options);
         }
-        if let Some(provider_obj) = storage_options_provider {
+
-            let provider = py_object_to_storage_options_provider(provider_obj)?;
+        // Auto-create storage options provider from namespace_client
+        if let Some(ns_obj) = namespace_client {
+            let ns_client = extract_namespace_arc(py, ns_obj)?;
+            // Create table_id by combining namespace_path with table name
+            let mut table_id = ns_path;
+            table_id.push(name);
+            let provider = create_namespace_storage_options_provider(ns_client.clone(), table_id);
             builder = builder.storage_options_provider(provider);
+            builder = builder.namespace_client(ns_client);
         }
+
         if let Some(index_cache_size) = index_cache_size {
             builder = builder.index_cache_size(index_cache_size);
         }
         if let Some(location) = location {
             builder = builder.location(location);
         }
+        // Pass managed_versioning if provided to avoid redundant describe_table call
+        if let Some(enabled) = managed_versioning {
+            builder = builder.managed_versioning(enabled);
+        }
 
         future_into_py(self_.py(), async move {
             let table = builder.execute().await.infer_error()?;
@@ -216,12 +254,12 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (target_table_name, source_uri, target_namespace=vec![], source_version=None, source_tag=None, is_shallow=true))]
+    #[pyo3(signature = (target_table_name, source_uri, target_namespace_path=None, source_version=None, source_tag=None, is_shallow=true))]
     pub fn clone_table(
         self_: PyRef<'_, Self>,
         target_table_name: String,
         source_uri: String,
-        target_namespace: Vec<String>,
+        target_namespace_path: Option<Vec<String>>,
         source_version: Option<u64>,
         source_tag: Option<String>,
         is_shallow: bool,
@@ -229,7 +267,7 @@ impl Connection {
         let inner = self_.get_inner()?.clone();
 
         let mut builder = inner.clone_table(target_table_name, source_uri);
-        builder = builder.target_namespace(target_namespace);
+        builder = builder.target_namespace(target_namespace_path.unwrap_or_default());
         if let Some(version) = source_version {
             builder = builder.source_version(version);
         }
@@ -244,52 +282,56 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (cur_name, new_name, cur_namespace=vec![], new_namespace=vec![]))]
+    #[pyo3(signature = (cur_name, new_name, cur_namespace_path=None, new_namespace_path=None))]
     pub fn rename_table(
         self_: PyRef<'_, Self>,
         cur_name: String,
         new_name: String,
-        cur_namespace: Vec<String>,
+        cur_namespace_path: Option<Vec<String>>,
-        new_namespace: Vec<String>,
+        new_namespace_path: Option<Vec<String>>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let cur_ns_path = cur_namespace_path.unwrap_or_default();
+        let new_ns_path = new_namespace_path.unwrap_or_default();
         future_into_py(self_.py(), async move {
             inner
-                .rename_table(cur_name, new_name, &cur_namespace, &new_namespace)
+                .rename_table(cur_name, new_name, &cur_ns_path, &new_ns_path)
                 .await
                 .infer_error()
         })
     }
 
-    #[pyo3(signature = (name, namespace=vec![]))]
+    #[pyo3(signature = (name, namespace_path=None))]
     pub fn drop_table(
         self_: PyRef<'_, Self>,
         name: String,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let ns_path = namespace_path.unwrap_or_default();
         future_into_py(self_.py(), async move {
-            inner.drop_table(name, &namespace).await.infer_error()
+            inner.drop_table(name, &ns_path).await.infer_error()
         })
     }
 
-    #[pyo3(signature = (namespace=vec![],))]
+    #[pyo3(signature = (namespace_path=None,))]
     pub fn drop_all_tables(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
+        let ns_path = namespace_path.unwrap_or_default();
         future_into_py(self_.py(), async move {
-            inner.drop_all_tables(&namespace).await.infer_error()
+            inner.drop_all_tables(&ns_path).await.infer_error()
         })
     }
 
     // Namespace management methods
 
-    #[pyo3(signature = (namespace=vec![], page_token=None, limit=None))]
+    #[pyo3(signature = (namespace_path=None, page_token=None, limit=None))]
     pub fn list_namespaces(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         page_token: Option<String>,
         limit: Option<u32>,
     ) -> PyResult<Bound<'_, PyAny>> {
@@ -298,11 +340,7 @@ impl Connection {
         future_into_py(py, async move {
             use lance_namespace::models::ListNamespacesRequest;
             let request = ListNamespacesRequest {
-                id: if namespace.is_empty() {
+                id: namespace_path,
-                    None
-                } else {
-                    Some(namespace)
-                },
                 page_token,
                 limit: limit.map(|l| l as i32),
                 ..Default::default()
@@ -317,10 +355,10 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (namespace, mode=None, properties=None))]
+    #[pyo3(signature = (namespace_path, mode=None, properties=None))]
     pub fn create_namespace(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Vec<String>,
         mode: Option<String>,
         properties: Option<std::collections::HashMap<String, String>>,
     ) -> PyResult<Bound<'_, PyAny>> {
@@ -336,11 +374,7 @@ impl Connection {
                 _ => None,
             });
             let request = CreateNamespaceRequest {
-                id: if namespace.is_empty() {
+                id: Some(namespace_path),
-                    None
-                } else {
-                    Some(namespace)
-                },
                 mode: mode_str,
                 properties,
                 ..Default::default()
@@ -354,10 +388,10 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (namespace, mode=None, behavior=None))]
+    #[pyo3(signature = (namespace_path, mode=None, behavior=None))]
     pub fn drop_namespace(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Vec<String>,
         mode: Option<String>,
         behavior: Option<String>,
     ) -> PyResult<Bound<'_, PyAny>> {
@@ -377,11 +411,7 @@ impl Connection {
                 _ => None,
             });
             let request = DropNamespaceRequest {
-                id: if namespace.is_empty() {
+                id: Some(namespace_path),
-                    None
-                } else {
-                    Some(namespace)
-                },
                 mode: mode_str,
                 behavior: behavior_str,
                 ..Default::default()
@@ -396,21 +426,17 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (namespace,))]
+    #[pyo3(signature = (namespace_path,))]
     pub fn describe_namespace(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Vec<String>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.get_inner()?.clone();
         let py = self_.py();
         future_into_py(py, async move {
             use lance_namespace::models::DescribeNamespaceRequest;
             let request = DescribeNamespaceRequest {
-                id: if namespace.is_empty() {
+                id: Some(namespace_path),
-                    None
-                } else {
-                    Some(namespace)
-                },
                 ..Default::default()
             };
             let response = inner.describe_namespace(request).await.infer_error()?;
@@ -422,10 +448,10 @@ impl Connection {
         })
     }
 
-    #[pyo3(signature = (namespace=vec![], page_token=None, limit=None))]
+    #[pyo3(signature = (namespace_path=None, page_token=None, limit=None))]
     pub fn list_tables(
         self_: PyRef<'_, Self>,
-        namespace: Vec<String>,
+        namespace_path: Option<Vec<String>>,
         page_token: Option<String>,
         limit: Option<u32>,
     ) -> PyResult<Bound<'_, PyAny>> {
@@ -434,11 +460,7 @@ impl Connection {
         future_into_py(py, async move {
             use lance_namespace::models::ListTablesRequest;
             let request = ListTablesRequest {
-                id: if namespace.is_empty() {
+                id: namespace_path,
-                    None
-                } else {
-                    Some(namespace)
-                },
                 page_token,
                 limit: limit.map(|l| l as i32),
                 ..Default::default()

python/src/error.rs

@@ -2,10 +2,10 @@
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 use pyo3::{
+    PyErr, PyResult, Python,
     exceptions::{PyIOError, PyNotImplementedError, PyOSError, PyRuntimeError, PyValueError},
     intern,
     types::{PyAnyMethods, PyNone},
-    PyErr, PyResult, Python,
 };
 
 use lancedb::error::Error as LanceError;

python/src/expr.rs

@@ -0,0 +1,175 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! PyO3 bindings for the LanceDB expression builder API.
+//!
+//! This module exposes [`PyExpr`] and helper free functions so Python can
+//! build type-safe filter / projection expressions that map directly to
+//! DataFusion [`Expr`] nodes, bypassing SQL string parsing.
+
+use arrow::{datatypes::DataType, pyarrow::PyArrowType};
+use lancedb::expr::{DfExpr, col as ldb_col, contains, expr_cast, lit as df_lit, lower, upper};
+use pyo3::{Bound, PyAny, PyResult, exceptions::PyValueError, prelude::*, pyfunction};
+
+/// A type-safe DataFusion expression.
+///
+/// Instances are constructed via the free functions [`expr_col`] and
+/// [`expr_lit`] and combined with the methods on this struct.  On the Python
+/// side a thin wrapper class (`lancedb.expr.Expr`) delegates to these methods
+/// and adds Python operator overloads.
+#[pyclass(name = "PyExpr")]
+#[derive(Clone)]
+pub struct PyExpr(pub DfExpr);
+
+#[pymethods]
+impl PyExpr {
+    // ── comparisons ──────────────────────────────────────────────────────────
+
+    fn eq(&self, other: &Self) -> Self {
+        Self(self.0.clone().eq(other.0.clone()))
+    }
+
+    fn ne(&self, other: &Self) -> Self {
+        Self(self.0.clone().not_eq(other.0.clone()))
+    }
+
+    fn lt(&self, other: &Self) -> Self {
+        Self(self.0.clone().lt(other.0.clone()))
+    }
+
+    fn lte(&self, other: &Self) -> Self {
+        Self(self.0.clone().lt_eq(other.0.clone()))
+    }
+
+    fn gt(&self, other: &Self) -> Self {
+        Self(self.0.clone().gt(other.0.clone()))
+    }
+
+    fn gte(&self, other: &Self) -> Self {
+        Self(self.0.clone().gt_eq(other.0.clone()))
+    }
+
+    // ── logical ──────────────────────────────────────────────────────────────
+
+    fn and_(&self, other: &Self) -> Self {
+        Self(self.0.clone().and(other.0.clone()))
+    }
+
+    fn or_(&self, other: &Self) -> Self {
+        Self(self.0.clone().or(other.0.clone()))
+    }
+
+    fn not_(&self) -> Self {
+        use std::ops::Not;
+        Self(self.0.clone().not())
+    }
+
+    // ── arithmetic ───────────────────────────────────────────────────────────
+
+    fn add(&self, other: &Self) -> Self {
+        use std::ops::Add;
+        Self(self.0.clone().add(other.0.clone()))
+    }
+
+    fn sub(&self, other: &Self) -> Self {
+        use std::ops::Sub;
+        Self(self.0.clone().sub(other.0.clone()))
+    }
+
+    fn mul(&self, other: &Self) -> Self {
+        use std::ops::Mul;
+        Self(self.0.clone().mul(other.0.clone()))
+    }
+
+    fn div(&self, other: &Self) -> Self {
+        use std::ops::Div;
+        Self(self.0.clone().div(other.0.clone()))
+    }
+
+    // ── string functions ─────────────────────────────────────────────────────
+
+    /// Convert string column to lowercase.
+    fn lower(&self) -> Self {
+        Self(lower(self.0.clone()))
+    }
+
+    /// Convert string column to uppercase.
+    fn upper(&self) -> Self {
+        Self(upper(self.0.clone()))
+    }
+
+    /// Test whether the string contains `substr`.
+    fn contains(&self, substr: &Self) -> Self {
+        Self(contains(self.0.clone(), substr.0.clone()))
+    }
+
+    // ── type cast ────────────────────────────────────────────────────────────
+
+    /// Cast the expression to `data_type`.
+    ///
+    /// `data_type` must be a PyArrow `DataType` (e.g. `pa.int32()`).
+    /// On the Python side, `lancedb.expr.Expr.cast` also accepts type name
+    /// strings via `pa.lib.ensure_type` before forwarding here.
+    fn cast(&self, data_type: PyArrowType<DataType>) -> Self {
+        Self(expr_cast(self.0.clone(), data_type.0))
+    }
+
+    // ── utilities ────────────────────────────────────────────────────────────
+
+    /// 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()))
+    }
+
+    fn __repr__(&self) -> PyResult<String> {
+        let sql =
+            lancedb::expr::expr_to_sql_string(&self.0).unwrap_or_else(|_| "<expr>".to_string());
+        Ok(format!("PyExpr({})", sql))
+    }
+}
+
+// ── free functions ────────────────────────────────────────────────────────────
+
+/// Create a column reference expression.
+///
+/// The column name is preserved exactly as given (case-sensitive), so
+/// `col("firstName")` correctly references a field named `firstName`.
+#[pyfunction]
+pub fn expr_col(name: &str) -> PyExpr {
+    PyExpr(ldb_col(name))
+}
+
+/// Create a literal value expression.
+///
+/// Supported Python types: `bool`, `int`, `float`, `str`.
+#[pyfunction]
+pub fn expr_lit(value: Bound<'_, PyAny>) -> PyResult<PyExpr> {
+    // bool must be checked before int because bool is a subclass of int in Python
+    if let Ok(b) = value.extract::<bool>() {
+        return Ok(PyExpr(df_lit(b)));
+    }
+    if let Ok(i) = value.extract::<i64>() {
+        return Ok(PyExpr(df_lit(i)));
+    }
+    if let Ok(f) = value.extract::<f64>() {
+        return Ok(PyExpr(df_lit(f)));
+    }
+    if let Ok(s) = value.extract::<String>() {
+        return Ok(PyExpr(df_lit(s)));
+    }
+    Err(PyValueError::new_err(format!(
+        "unsupported literal type: {}. Supported: bool, int, float, str",
+        value.get_type().name()?
+    )))
+}
+
+/// Call an arbitrary registered SQL function by name.
+///
+/// See `lancedb::expr::func` for the list of supported function names.
+#[pyfunction]
+pub fn expr_func(name: &str, args: Vec<PyExpr>) -> PyResult<PyExpr> {
+    let df_args: Vec<DfExpr> = args.into_iter().map(|e| e.0).collect();
+    lancedb::expr::func(name, df_args)
+        .map(PyExpr)
+        .map_err(|e| PyValueError::new_err(e.to_string()))
+}

python/src/index.rs

@@ -3,17 +3,17 @@
 
 use lancedb::index::vector::{IvfFlatIndexBuilder, IvfRqIndexBuilder, IvfSqIndexBuilder};
 use lancedb::index::{
+    Index as LanceDbIndex,
     scalar::{BTreeIndexBuilder, FtsIndexBuilder},
     vector::{IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder},
-    Index as LanceDbIndex,
 };
-use pyo3::types::PyStringMethods;
 use pyo3::IntoPyObject;
+use pyo3::types::PyStringMethods;
 use pyo3::{
+    Bound, FromPyObject, PyAny, PyResult, Python,
     exceptions::{PyKeyError, PyValueError},
     intern, pyclass, pymethods,
     types::PyAnyMethods,
-    Bound, FromPyObject, PyAny, PyResult, Python,
 };
 
 use crate::util::parse_distance_type;
@@ -41,7 +41,12 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                 let inner_opts = FtsIndexBuilder::default()
                     .base_tokenizer(params.base_tokenizer)
                     .language(&params.language)
-                    .map_err(|_| PyValueError::new_err(format!("LanceDB does not support the requested language: '{}'", params.language)))?
+                    .map_err(|_| {
+                        PyValueError::new_err(format!(
+                            "LanceDB does not support the requested language: '{}'",
+                            params.language
+                        ))
+                    })?
                     .with_position(params.with_position)
                     .lower_case(params.lower_case)
                     .max_token_length(params.max_token_length)
@@ -52,7 +57,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     .ngram_max_length(params.ngram_max_length)
                     .ngram_prefix_only(params.prefix_only);
                 Ok(LanceDbIndex::FTS(inner_opts))
-            },
+            }
             "IvfFlat" => {
                 let params = source.extract::<IvfFlatParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -64,10 +69,11 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     ivf_flat_builder = ivf_flat_builder.num_partitions(num_partitions);
                 }
                 if let Some(target_partition_size) = params.target_partition_size {
-                    ivf_flat_builder = ivf_flat_builder.target_partition_size(target_partition_size);
+                    ivf_flat_builder =
+                        ivf_flat_builder.target_partition_size(target_partition_size);
                 }
                 Ok(LanceDbIndex::IvfFlat(ivf_flat_builder))
-            },
+            }
             "IvfPq" => {
                 let params = source.extract::<IvfPqParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -86,7 +92,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     ivf_pq_builder = ivf_pq_builder.num_sub_vectors(num_sub_vectors);
                 }
                 Ok(LanceDbIndex::IvfPq(ivf_pq_builder))
-            },
+            }
             "IvfSq" => {
                 let params = source.extract::<IvfSqParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -101,7 +107,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     ivf_sq_builder = ivf_sq_builder.target_partition_size(target_partition_size);
                 }
                 Ok(LanceDbIndex::IvfSq(ivf_sq_builder))
-            },
+            }
             "IvfRq" => {
                 let params = source.extract::<IvfRqParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -117,7 +123,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     ivf_rq_builder = ivf_rq_builder.target_partition_size(target_partition_size);
                 }
                 Ok(LanceDbIndex::IvfRq(ivf_rq_builder))
-            },
+            }
             "HnswPq" => {
                 let params = source.extract::<IvfHnswPqParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -138,7 +144,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     hnsw_pq_builder = hnsw_pq_builder.num_sub_vectors(num_sub_vectors);
                 }
                 Ok(LanceDbIndex::IvfHnswPq(hnsw_pq_builder))
-            },
+            }
             "HnswSq" => {
                 let params = source.extract::<IvfHnswSqParams>()?;
                 let distance_type = parse_distance_type(params.distance_type)?;
@@ -155,7 +161,7 @@ pub fn extract_index_params(source: &Option<Bound<'_, PyAny>>) -> PyResult<Lance
                     hnsw_sq_builder = hnsw_sq_builder.target_partition_size(target_partition_size);
                 }
                 Ok(LanceDbIndex::IvfHnswSq(hnsw_sq_builder))
-            },
+            }
             not_supported => Err(PyValueError::new_err(format!(
                 "Invalid index type '{}'.  Must be one of BTree, Bitmap, LabelList, FTS, IvfPq, IvfSq, IvfHnswPq, or IvfHnswSq",
                 not_supported

python/src/lib.rs

@@ -2,14 +2,15 @@
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 use arrow::RecordBatchStream;
-use connection::{connect, Connection};
+use connection::{Connection, connect};
 use env_logger::Env;
+use expr::{PyExpr, expr_col, expr_func, expr_lit};
 use index::IndexConfig;
 use permutation::{PyAsyncPermutationBuilder, PyPermutationReader};
 use pyo3::{
-    pymodule,
+    Bound, PyResult, Python, pymodule,
     types::{PyModule, PyModuleMethods},
-    wrap_pyfunction, Bound, PyResult, Python,
+    wrap_pyfunction,
 };
 use query::{FTSQuery, HybridQuery, Query, VectorQuery};
 use session::Session;
@@ -21,12 +22,13 @@ use table::{
 pub mod arrow;
 pub mod connection;
 pub mod error;
+pub mod expr;
 pub mod header;
 pub mod index;
+pub mod namespace;
 pub mod permutation;
 pub mod query;
 pub mod session;
-pub mod storage_options;
 pub mod table;
 pub mod util;
 
@@ -54,10 +56,14 @@ pub fn _lancedb(_py: Python, m: &Bound<'_, PyModule>) -> PyResult<()> {
     m.add_class::<UpdateResult>()?;
     m.add_class::<PyAsyncPermutationBuilder>()?;
     m.add_class::<PyPermutationReader>()?;
+    m.add_class::<PyExpr>()?;
     m.add_function(wrap_pyfunction!(connect, m)?)?;
     m.add_function(wrap_pyfunction!(permutation::async_permutation_builder, m)?)?;
     m.add_function(wrap_pyfunction!(util::validate_table_name, m)?)?;
     m.add_function(wrap_pyfunction!(query::fts_query_to_json, m)?)?;
+    m.add_function(wrap_pyfunction!(expr_col, m)?)?;
+    m.add_function(wrap_pyfunction!(expr_lit, m)?)?;
+    m.add_function(wrap_pyfunction!(expr_func, m)?)?;
     m.add("__version__", env!("CARGO_PKG_VERSION"))?;
     Ok(())
 }

python/src/namespace.rs

@@ -0,0 +1,715 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! Namespace utilities for Python bindings
+
+use std::collections::HashMap;
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use bytes::Bytes;
+use lance_io::object_store::{LanceNamespaceStorageOptionsProvider, StorageOptionsProvider};
+use lance_namespace::LanceNamespace as LanceNamespaceTrait;
+use lance_namespace::models::*;
+use pyo3::prelude::*;
+use pyo3::types::PyDict;
+
+/// Wrapper that allows any Python object implementing LanceNamespace protocol
+/// to be used as a Rust LanceNamespace.
+///
+/// This is similar to PyLanceNamespace in lance's Python bindings - it wraps a Python
+/// object and calls back into Python when namespace methods are invoked.
+pub struct PyLanceNamespace {
+    py_namespace: Arc<Py<PyAny>>,
+    namespace_id: String,
+}
+
+impl PyLanceNamespace {
+    /// Create a new PyLanceNamespace wrapper around a Python namespace object.
+    pub fn new(_py: Python<'_>, py_namespace: &Bound<'_, PyAny>) -> PyResult<Self> {
+        let namespace_id = py_namespace
+            .call_method0("namespace_id")?
+            .extract::<String>()?;
+
+        Ok(Self {
+            py_namespace: Arc::new(py_namespace.clone().unbind()),
+            namespace_id,
+        })
+    }
+
+    /// Create an Arc<dyn LanceNamespace> from a Python namespace object.
+    pub fn create_arc(
+        py: Python<'_>,
+        py_namespace: &Bound<'_, PyAny>,
+    ) -> PyResult<Arc<dyn LanceNamespaceTrait>> {
+        let wrapper = Self::new(py, py_namespace)?;
+        Ok(Arc::new(wrapper))
+    }
+}
+
+impl std::fmt::Debug for PyLanceNamespace {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "PyLanceNamespace {{ id: {} }}", self.namespace_id)
+    }
+}
+
+/// Get or create the DictWithModelDump class in Python.
+/// This class acts like a dict but also has model_dump() method.
+/// This allows it to work with both:
+/// - depythonize (which expects a dict/Mapping)
+/// - Python code that calls .model_dump() (like DirectoryNamespace wrapper)
+fn get_dict_with_model_dump_class(py: Python<'_>) -> PyResult<Bound<'_, PyAny>> {
+    // Use a module-level cache via __builtins__
+    let builtins = py.import("builtins")?;
+    if builtins.hasattr("_DictWithModelDump")? {
+        return builtins.getattr("_DictWithModelDump");
+    }
+
+    // Create the class using exec
+    let locals = PyDict::new(py);
+    py.run(
+        c"class DictWithModelDump(dict):
+    def model_dump(self):
+        return dict(self)",
+        None,
+        Some(&locals),
+    )?;
+    let class = locals.get_item("DictWithModelDump")?.ok_or_else(|| {
+        pyo3::exceptions::PyRuntimeError::new_err("Failed to create DictWithModelDump class")
+    })?;
+
+    // Cache it
+    builtins.setattr("_DictWithModelDump", &class)?;
+    Ok(class)
+}
+
+/// Helper to call a Python namespace method with JSON serialization.
+/// For methods that take a request and return a response.
+/// Uses DictWithModelDump to pass a dict that also has model_dump() method,
+/// making it compatible with both depythonize and Python wrappers.
+async fn call_py_method<Req, Resp>(
+    py_namespace: Arc<Py<PyAny>>,
+    method_name: &'static str,
+    request: Req,
+) -> lance_core::Result<Resp>
+where
+    Req: serde::Serialize + Send + 'static,
+    Resp: serde::de::DeserializeOwned + Send + 'static,
+{
+    let request_json = serde_json::to_string(&request).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to serialize request for {}: {}",
+            method_name, e
+        ))
+    })?;
+
+    let response_json = tokio::task::spawn_blocking(move || {
+        Python::attach(|py| {
+            let json_module = py.import("json")?;
+            let request_dict = json_module.call_method1("loads", (&request_json,))?;
+
+            // Wrap dict in DictWithModelDump so it works with both depythonize and .model_dump()
+            let dict_class = get_dict_with_model_dump_class(py)?;
+            let request_arg = dict_class.call1((request_dict,))?;
+
+            // Call the Python method
+            let result = py_namespace.call_method1(py, method_name, (request_arg,))?;
+
+            // Convert response to dict, then to JSON
+            // Pydantic models have model_dump() method
+            let result_dict = if result.bind(py).hasattr("model_dump")? {
+                result.call_method0(py, "model_dump")?
+            } else {
+                result
+            };
+            let response_json: String = json_module
+                .call_method1("dumps", (result_dict,))?
+                .extract()?;
+            Ok::<_, PyErr>(response_json)
+        })
+    })
+    .await
+    .map_err(|e| lance_core::Error::io(format!("Task join error for {}: {}", method_name, e)))?
+    .map_err(|e: PyErr| lance_core::Error::io(format!("Python error in {}: {}", method_name, e)))?;
+
+    serde_json::from_str(&response_json).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to deserialize response from {}: {}",
+            method_name, e
+        ))
+    })
+}
+
+/// Helper for methods that return () on success
+async fn call_py_method_unit<Req>(
+    py_namespace: Arc<Py<PyAny>>,
+    method_name: &'static str,
+    request: Req,
+) -> lance_core::Result<()>
+where
+    Req: serde::Serialize + Send + 'static,
+{
+    let request_json = serde_json::to_string(&request).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to serialize request for {}: {}",
+            method_name, e
+        ))
+    })?;
+
+    tokio::task::spawn_blocking(move || {
+        Python::attach(|py| {
+            let json_module = py.import("json")?;
+            let request_dict = json_module.call_method1("loads", (&request_json,))?;
+
+            // Wrap dict in DictWithModelDump
+            let dict_class = get_dict_with_model_dump_class(py)?;
+            let request_arg = dict_class.call1((request_dict,))?;
+
+            // Call the Python method
+            py_namespace.call_method1(py, method_name, (request_arg,))?;
+            Ok::<_, PyErr>(())
+        })
+    })
+    .await
+    .map_err(|e| lance_core::Error::io(format!("Task join error for {}: {}", method_name, e)))?
+    .map_err(|e: PyErr| lance_core::Error::io(format!("Python error in {}: {}", method_name, e)))
+}
+
+/// Helper for methods that return a primitive type
+async fn call_py_method_primitive<Req, Resp>(
+    py_namespace: Arc<Py<PyAny>>,
+    method_name: &'static str,
+    request: Req,
+) -> lance_core::Result<Resp>
+where
+    Req: serde::Serialize + Send + 'static,
+    Resp: for<'py> pyo3::FromPyObject<'py> + Send + 'static,
+{
+    let request_json = serde_json::to_string(&request).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to serialize request for {}: {}",
+            method_name, e
+        ))
+    })?;
+
+    tokio::task::spawn_blocking(move || {
+        Python::attach(|py| {
+            let json_module = py.import("json")?;
+            let request_dict = json_module.call_method1("loads", (&request_json,))?;
+
+            // Wrap dict in DictWithModelDump
+            let dict_class = get_dict_with_model_dump_class(py)?;
+            let request_arg = dict_class.call1((request_dict,))?;
+
+            // Call the Python method
+            let result = py_namespace.call_method1(py, method_name, (request_arg,))?;
+            let value: Resp = result.extract(py)?;
+            Ok::<_, PyErr>(value)
+        })
+    })
+    .await
+    .map_err(|e| lance_core::Error::io(format!("Task join error for {}: {}", method_name, e)))?
+    .map_err(|e: PyErr| lance_core::Error::io(format!("Python error in {}: {}", method_name, e)))
+}
+
+/// Helper for methods that return Bytes
+async fn call_py_method_bytes<Req>(
+    py_namespace: Arc<Py<PyAny>>,
+    method_name: &'static str,
+    request: Req,
+) -> lance_core::Result<Bytes>
+where
+    Req: serde::Serialize + Send + 'static,
+{
+    let request_json = serde_json::to_string(&request).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to serialize request for {}: {}",
+            method_name, e
+        ))
+    })?;
+
+    tokio::task::spawn_blocking(move || {
+        Python::attach(|py| {
+            let json_module = py.import("json")?;
+            let request_dict = json_module.call_method1("loads", (&request_json,))?;
+
+            // Wrap dict in DictWithModelDump
+            let dict_class = get_dict_with_model_dump_class(py)?;
+            let request_arg = dict_class.call1((request_dict,))?;
+
+            // Call the Python method
+            let result = py_namespace.call_method1(py, method_name, (request_arg,))?;
+            let bytes_data: Vec<u8> = result.extract(py)?;
+            Ok::<_, PyErr>(Bytes::from(bytes_data))
+        })
+    })
+    .await
+    .map_err(|e| lance_core::Error::io(format!("Task join error for {}: {}", method_name, e)))?
+    .map_err(|e: PyErr| lance_core::Error::io(format!("Python error in {}: {}", method_name, e)))
+}
+
+/// Helper for methods that take request + data and return a response
+async fn call_py_method_with_data<Req, Resp>(
+    py_namespace: Arc<Py<PyAny>>,
+    method_name: &'static str,
+    request: Req,
+    data: Bytes,
+) -> lance_core::Result<Resp>
+where
+    Req: serde::Serialize + Send + 'static,
+    Resp: serde::de::DeserializeOwned + Send + 'static,
+{
+    let request_json = serde_json::to_string(&request).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to serialize request for {}: {}",
+            method_name, e
+        ))
+    })?;
+
+    let response_json = tokio::task::spawn_blocking(move || {
+        Python::attach(|py| {
+            let json_module = py.import("json")?;
+            let request_dict = json_module.call_method1("loads", (&request_json,))?;
+
+            // Wrap dict in DictWithModelDump
+            let dict_class = get_dict_with_model_dump_class(py)?;
+            let request_arg = dict_class.call1((request_dict,))?;
+
+            // Pass request and bytes to Python method
+            let py_bytes = pyo3::types::PyBytes::new(py, &data);
+            let result = py_namespace.call_method1(py, method_name, (request_arg, py_bytes))?;
+
+            // Convert response dict to JSON
+            let response_json: String = json_module.call_method1("dumps", (result,))?.extract()?;
+            Ok::<_, PyErr>(response_json)
+        })
+    })
+    .await
+    .map_err(|e| lance_core::Error::io(format!("Task join error for {}: {}", method_name, e)))?
+    .map_err(|e: PyErr| lance_core::Error::io(format!("Python error in {}: {}", method_name, e)))?;
+
+    serde_json::from_str(&response_json).map_err(|e| {
+        lance_core::Error::io(format!(
+            "Failed to deserialize response from {}: {}",
+            method_name, e
+        ))
+    })
+}
+
+#[async_trait]
+impl LanceNamespaceTrait for PyLanceNamespace {
+    fn namespace_id(&self) -> String {
+        self.namespace_id.clone()
+    }
+
+    async fn list_namespaces(
+        &self,
+        request: ListNamespacesRequest,
+    ) -> lance_core::Result<ListNamespacesResponse> {
+        call_py_method(self.py_namespace.clone(), "list_namespaces", request).await
+    }
+
+    async fn describe_namespace(
+        &self,
+        request: DescribeNamespaceRequest,
+    ) -> lance_core::Result<DescribeNamespaceResponse> {
+        call_py_method(self.py_namespace.clone(), "describe_namespace", request).await
+    }
+
+    async fn create_namespace(
+        &self,
+        request: CreateNamespaceRequest,
+    ) -> lance_core::Result<CreateNamespaceResponse> {
+        call_py_method(self.py_namespace.clone(), "create_namespace", request).await
+    }
+
+    async fn drop_namespace(
+        &self,
+        request: DropNamespaceRequest,
+    ) -> lance_core::Result<DropNamespaceResponse> {
+        call_py_method(self.py_namespace.clone(), "drop_namespace", request).await
+    }
+
+    async fn namespace_exists(&self, request: NamespaceExistsRequest) -> lance_core::Result<()> {
+        call_py_method_unit(self.py_namespace.clone(), "namespace_exists", request).await
+    }
+
+    async fn list_tables(
+        &self,
+        request: ListTablesRequest,
+    ) -> lance_core::Result<ListTablesResponse> {
+        call_py_method(self.py_namespace.clone(), "list_tables", request).await
+    }
+
+    async fn describe_table(
+        &self,
+        request: DescribeTableRequest,
+    ) -> lance_core::Result<DescribeTableResponse> {
+        call_py_method(self.py_namespace.clone(), "describe_table", request).await
+    }
+
+    async fn register_table(
+        &self,
+        request: RegisterTableRequest,
+    ) -> lance_core::Result<RegisterTableResponse> {
+        call_py_method(self.py_namespace.clone(), "register_table", request).await
+    }
+
+    async fn table_exists(&self, request: TableExistsRequest) -> lance_core::Result<()> {
+        call_py_method_unit(self.py_namespace.clone(), "table_exists", request).await
+    }
+
+    async fn drop_table(&self, request: DropTableRequest) -> lance_core::Result<DropTableResponse> {
+        call_py_method(self.py_namespace.clone(), "drop_table", request).await
+    }
+
+    async fn deregister_table(
+        &self,
+        request: DeregisterTableRequest,
+    ) -> lance_core::Result<DeregisterTableResponse> {
+        call_py_method(self.py_namespace.clone(), "deregister_table", request).await
+    }
+
+    async fn count_table_rows(&self, request: CountTableRowsRequest) -> lance_core::Result<i64> {
+        call_py_method_primitive(self.py_namespace.clone(), "count_table_rows", request).await
+    }
+
+    async fn create_table(
+        &self,
+        request: CreateTableRequest,
+        request_data: Bytes,
+    ) -> lance_core::Result<CreateTableResponse> {
+        call_py_method_with_data(
+            self.py_namespace.clone(),
+            "create_table",
+            request,
+            request_data,
+        )
+        .await
+    }
+
+    async fn declare_table(
+        &self,
+        request: DeclareTableRequest,
+    ) -> lance_core::Result<DeclareTableResponse> {
+        call_py_method(self.py_namespace.clone(), "declare_table", request).await
+    }
+
+    async fn insert_into_table(
+        &self,
+        request: InsertIntoTableRequest,
+        request_data: Bytes,
+    ) -> lance_core::Result<InsertIntoTableResponse> {
+        call_py_method_with_data(
+            self.py_namespace.clone(),
+            "insert_into_table",
+            request,
+            request_data,
+        )
+        .await
+    }
+
+    async fn merge_insert_into_table(
+        &self,
+        request: MergeInsertIntoTableRequest,
+        request_data: Bytes,
+    ) -> lance_core::Result<MergeInsertIntoTableResponse> {
+        call_py_method_with_data(
+            self.py_namespace.clone(),
+            "merge_insert_into_table",
+            request,
+            request_data,
+        )
+        .await
+    }
+
+    async fn update_table(
+        &self,
+        request: UpdateTableRequest,
+    ) -> lance_core::Result<UpdateTableResponse> {
+        call_py_method(self.py_namespace.clone(), "update_table", request).await
+    }
+
+    async fn delete_from_table(
+        &self,
+        request: DeleteFromTableRequest,
+    ) -> lance_core::Result<DeleteFromTableResponse> {
+        call_py_method(self.py_namespace.clone(), "delete_from_table", request).await
+    }
+
+    async fn query_table(&self, request: QueryTableRequest) -> lance_core::Result<Bytes> {
+        call_py_method_bytes(self.py_namespace.clone(), "query_table", request).await
+    }
+
+    async fn create_table_index(
+        &self,
+        request: CreateTableIndexRequest,
+    ) -> lance_core::Result<CreateTableIndexResponse> {
+        call_py_method(self.py_namespace.clone(), "create_table_index", request).await
+    }
+
+    async fn list_table_indices(
+        &self,
+        request: ListTableIndicesRequest,
+    ) -> lance_core::Result<ListTableIndicesResponse> {
+        call_py_method(self.py_namespace.clone(), "list_table_indices", request).await
+    }
+
+    async fn describe_table_index_stats(
+        &self,
+        request: DescribeTableIndexStatsRequest,
+    ) -> lance_core::Result<DescribeTableIndexStatsResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "describe_table_index_stats",
+            request,
+        )
+        .await
+    }
+
+    async fn describe_transaction(
+        &self,
+        request: DescribeTransactionRequest,
+    ) -> lance_core::Result<DescribeTransactionResponse> {
+        call_py_method(self.py_namespace.clone(), "describe_transaction", request).await
+    }
+
+    async fn alter_transaction(
+        &self,
+        request: AlterTransactionRequest,
+    ) -> lance_core::Result<AlterTransactionResponse> {
+        call_py_method(self.py_namespace.clone(), "alter_transaction", request).await
+    }
+
+    async fn create_table_scalar_index(
+        &self,
+        request: CreateTableIndexRequest,
+    ) -> lance_core::Result<CreateTableScalarIndexResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "create_table_scalar_index",
+            request,
+        )
+        .await
+    }
+
+    async fn drop_table_index(
+        &self,
+        request: DropTableIndexRequest,
+    ) -> lance_core::Result<DropTableIndexResponse> {
+        call_py_method(self.py_namespace.clone(), "drop_table_index", request).await
+    }
+
+    async fn list_all_tables(
+        &self,
+        request: ListTablesRequest,
+    ) -> lance_core::Result<ListTablesResponse> {
+        call_py_method(self.py_namespace.clone(), "list_all_tables", request).await
+    }
+
+    async fn restore_table(
+        &self,
+        request: RestoreTableRequest,
+    ) -> lance_core::Result<RestoreTableResponse> {
+        call_py_method(self.py_namespace.clone(), "restore_table", request).await
+    }
+
+    async fn rename_table(
+        &self,
+        request: RenameTableRequest,
+    ) -> lance_core::Result<RenameTableResponse> {
+        call_py_method(self.py_namespace.clone(), "rename_table", request).await
+    }
+
+    async fn list_table_versions(
+        &self,
+        request: ListTableVersionsRequest,
+    ) -> lance_core::Result<ListTableVersionsResponse> {
+        call_py_method(self.py_namespace.clone(), "list_table_versions", request).await
+    }
+
+    async fn create_table_version(
+        &self,
+        request: CreateTableVersionRequest,
+    ) -> lance_core::Result<CreateTableVersionResponse> {
+        call_py_method(self.py_namespace.clone(), "create_table_version", request).await
+    }
+
+    async fn describe_table_version(
+        &self,
+        request: DescribeTableVersionRequest,
+    ) -> lance_core::Result<DescribeTableVersionResponse> {
+        call_py_method(self.py_namespace.clone(), "describe_table_version", request).await
+    }
+
+    async fn batch_delete_table_versions(
+        &self,
+        request: BatchDeleteTableVersionsRequest,
+    ) -> lance_core::Result<BatchDeleteTableVersionsResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "batch_delete_table_versions",
+            request,
+        )
+        .await
+    }
+
+    async fn update_table_schema_metadata(
+        &self,
+        request: UpdateTableSchemaMetadataRequest,
+    ) -> lance_core::Result<UpdateTableSchemaMetadataResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "update_table_schema_metadata",
+            request,
+        )
+        .await
+    }
+
+    async fn get_table_stats(
+        &self,
+        request: GetTableStatsRequest,
+    ) -> lance_core::Result<GetTableStatsResponse> {
+        call_py_method(self.py_namespace.clone(), "get_table_stats", request).await
+    }
+
+    async fn explain_table_query_plan(
+        &self,
+        request: ExplainTableQueryPlanRequest,
+    ) -> lance_core::Result<String> {
+        call_py_method_primitive(
+            self.py_namespace.clone(),
+            "explain_table_query_plan",
+            request,
+        )
+        .await
+    }
+
+    async fn analyze_table_query_plan(
+        &self,
+        request: AnalyzeTableQueryPlanRequest,
+    ) -> lance_core::Result<String> {
+        call_py_method_primitive(
+            self.py_namespace.clone(),
+            "analyze_table_query_plan",
+            request,
+        )
+        .await
+    }
+
+    async fn alter_table_add_columns(
+        &self,
+        request: AlterTableAddColumnsRequest,
+    ) -> lance_core::Result<AlterTableAddColumnsResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "alter_table_add_columns",
+            request,
+        )
+        .await
+    }
+
+    async fn alter_table_alter_columns(
+        &self,
+        request: AlterTableAlterColumnsRequest,
+    ) -> lance_core::Result<AlterTableAlterColumnsResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "alter_table_alter_columns",
+            request,
+        )
+        .await
+    }
+
+    async fn alter_table_drop_columns(
+        &self,
+        request: AlterTableDropColumnsRequest,
+    ) -> lance_core::Result<AlterTableDropColumnsResponse> {
+        call_py_method(
+            self.py_namespace.clone(),
+            "alter_table_drop_columns",
+            request,
+        )
+        .await
+    }
+
+    async fn list_table_tags(
+        &self,
+        request: ListTableTagsRequest,
+    ) -> lance_core::Result<ListTableTagsResponse> {
+        call_py_method(self.py_namespace.clone(), "list_table_tags", request).await
+    }
+
+    async fn create_table_tag(
+        &self,
+        request: CreateTableTagRequest,
+    ) -> lance_core::Result<CreateTableTagResponse> {
+        call_py_method(self.py_namespace.clone(), "create_table_tag", request).await
+    }
+
+    async fn delete_table_tag(
+        &self,
+        request: DeleteTableTagRequest,
+    ) -> lance_core::Result<DeleteTableTagResponse> {
+        call_py_method(self.py_namespace.clone(), "delete_table_tag", request).await
+    }
+
+    async fn update_table_tag(
+        &self,
+        request: UpdateTableTagRequest,
+    ) -> lance_core::Result<UpdateTableTagResponse> {
+        call_py_method(self.py_namespace.clone(), "update_table_tag", request).await
+    }
+
+    async fn get_table_tag_version(
+        &self,
+        request: GetTableTagVersionRequest,
+    ) -> lance_core::Result<GetTableTagVersionResponse> {
+        call_py_method(self.py_namespace.clone(), "get_table_tag_version", request).await
+    }
+}
+
+/// Convert Python dict to HashMap<String, String>
+#[allow(dead_code)]
+fn dict_to_hashmap(dict: &Bound<'_, PyDict>) -> PyResult<HashMap<String, String>> {
+    let mut map = HashMap::new();
+    for (key, value) in dict.iter() {
+        let key_str: String = key.extract()?;
+        let value_str: String = value.extract()?;
+        map.insert(key_str, value_str);
+    }
+    Ok(map)
+}
+
+/// Extract an Arc<dyn LanceNamespace> from a Python namespace object.
+///
+/// This function wraps any Python namespace object with PyLanceNamespace.
+/// The PyLanceNamespace wrapper uses DictWithModelDump to pass requests,
+/// which works with both:
+/// - Native namespaces (DirectoryNamespace, RestNamespace) that use depythonize (expects dict)
+/// - Custom Python implementations that call .model_dump() on the request
+pub fn extract_namespace_arc(
+    py: Python<'_>,
+    ns: Py<PyAny>,
+) -> PyResult<Arc<dyn LanceNamespaceTrait>> {
+    let ns_ref = ns.bind(py);
+    PyLanceNamespace::create_arc(py, ns_ref)
+}
+
+/// Create a LanceNamespaceStorageOptionsProvider from a namespace client and table ID.
+///
+/// This creates a Rust storage options provider that fetches credentials from the
+/// namespace's describe_table() method, enabling automatic credential refresh.
+///
+/// # Arguments
+/// * `namespace_client` - The namespace client (wrapped PyLanceNamespace)
+/// * `table_id` - Full table identifier (namespace_path + table_name)
+pub fn create_namespace_storage_options_provider(
+    namespace_client: Arc<dyn LanceNamespaceTrait>,
+    table_id: Vec<String>,
+) -> Arc<dyn StorageOptionsProvider> {
+    Arc::new(LanceNamespaceStorageOptionsProvider::new(
+        namespace_client,
+        table_id,
+    ))
+}

python/src/permutation.rs

@@ -16,10 +16,10 @@ use lancedb::{
     query::Select,
 };
 use pyo3::{
+    Bound, PyAny, PyRef, PyRefMut, PyResult, Python,
     exceptions::PyRuntimeError,
     pyclass, pymethods,
     types::{PyAnyMethods, PyDict, PyDictMethods, PyType},
-    Bound, PyAny, PyRef, PyRefMut, PyResult, Python,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 

python/src/query.rs

@@ -4,9 +4,9 @@
 use std::sync::Arc;
 use std::time::Duration;
 
-use arrow::array::make_array;
 use arrow::array::Array;
 use arrow::array::ArrayData;
+use arrow::array::make_array;
 use arrow::pyarrow::FromPyArrow;
 use arrow::pyarrow::IntoPyArrow;
 use arrow::pyarrow::ToPyArrow;
@@ -22,25 +22,23 @@ use lancedb::query::{
     VectorQuery as LanceDbVectorQuery,
 };
 use lancedb::table::AnyQuery;
-use pyo3::prelude::{PyAnyMethods, PyDictMethods};
-use pyo3::pyfunction;
-use pyo3::pymethods;
-use pyo3::types::PyList;
-use pyo3::types::{PyDict, PyString};
 use pyo3::Bound;
 use pyo3::IntoPyObject;
 use pyo3::PyAny;
 use pyo3::PyRef;
 use pyo3::PyResult;
 use pyo3::Python;
-use pyo3::{exceptions::PyRuntimeError, FromPyObject};
+use pyo3::prelude::{PyAnyMethods, PyDictMethods};
-use pyo3::{
+use pyo3::pyfunction;
-    exceptions::{PyNotImplementedError, PyValueError},
+use pyo3::pymethods;
-    intern,
+use pyo3::types::PyList;
-};
+use pyo3::types::{PyDict, PyString};
-use pyo3::{pyclass, PyErr};
+use pyo3::{FromPyObject, exceptions::PyRuntimeError};
+use pyo3::{PyErr, pyclass};
+use pyo3::{exceptions::PyValueError, intern};
 use pyo3_async_runtimes::tokio::future_into_py;
 
+use crate::expr::PyExpr;
 use crate::util::parse_distance_type;
 use crate::{arrow::RecordBatchStream, util::PyLanceDB};
 use crate::{error::PythonErrorExt, index::class_name};
@@ -316,6 +314,19 @@ impl<'py> IntoPyObject<'py> for PySelect {
             Select::All => Ok(py.None().into_bound(py).into_any()),
             Select::Columns(columns) => Ok(columns.into_pyobject(py)?.into_any()),
             Select::Dynamic(columns) => Ok(columns.into_pyobject(py)?.into_any()),
+            Select::Expr(pairs) => {
+                // Serialize DataFusion Expr -> SQL string so Python sees the same
+                // format as Select::Dynamic: a list of (name, sql_string) tuples.
+                let sql_pairs: PyResult<Vec<(String, String)>> = pairs
+                    .into_iter()
+                    .map(|(name, expr)| {
+                        lancedb::expr::expr_to_sql_string(&expr)
+                            .map(|sql| (name, sql))
+                            .map_err(|e| PyRuntimeError::new_err(e.to_string()))
+                    })
+                    .collect();
+                Ok(sql_pairs?.into_pyobject(py)?.into_any())
+            }
         }
     }
 }
@@ -331,9 +342,13 @@ impl<'py> IntoPyObject<'py> for PyQueryFilter {
 
     fn into_pyobject(self, py: pyo3::Python<'py>) -> PyResult<Self::Output> {
         match self.0 {
-            QueryFilter::Datafusion(_) => Err(PyNotImplementedError::new_err(
+            QueryFilter::Datafusion(expr) => {
-                "Datafusion filter has no conversion to Python",
+                // Serialize the DataFusion expression to a SQL string so that
-            )),
+                // callers (e.g. remote tables) see the same format as Sql.
+                let sql = lancedb::expr::expr_to_sql_string(&expr)
+                    .map_err(|e| PyRuntimeError::new_err(e.to_string()))?;
+                Ok(sql.into_pyobject(py)?.into_any())
+            }
             QueryFilter::Sql(sql) => Ok(sql.into_pyobject(py)?.into_any()),
             QueryFilter::Substrait(substrait) => Ok(substrait.into_pyobject(py)?.into_any()),
         }
@@ -357,10 +372,20 @@ impl Query {
         self.inner = self.inner.clone().only_if(predicate);
     }
 
+    pub fn where_expr(&mut self, expr: PyExpr) {
+        self.inner = self.inner.clone().only_if_expr(expr.0);
+    }
+
     pub fn select(&mut self, columns: Vec<(String, String)>) {
         self.inner = self.inner.clone().select(Select::dynamic(&columns));
     }
 
+    pub fn select_expr(&mut self, columns: Vec<(String, PyExpr)>) {
+        let pairs: Vec<(String, lancedb::expr::DfExpr)> =
+            columns.into_iter().map(|(name, e)| (name, e.0)).collect();
+        self.inner = self.inner.clone().select(Select::Expr(pairs));
+    }
+
     pub fn select_columns(&mut self, columns: Vec<String>) {
         self.inner = self.inner.clone().select(Select::columns(&columns));
     }
@@ -594,10 +619,20 @@ impl FTSQuery {
         self.inner = self.inner.clone().only_if(predicate);
     }
 
+    pub fn where_expr(&mut self, expr: PyExpr) {
+        self.inner = self.inner.clone().only_if_expr(expr.0);
+    }
+
     pub fn select(&mut self, columns: Vec<(String, String)>) {
         self.inner = self.inner.clone().select(Select::dynamic(&columns));
     }
 
+    pub fn select_expr(&mut self, columns: Vec<(String, PyExpr)>) {
+        let pairs: Vec<(String, lancedb::expr::DfExpr)> =
+            columns.into_iter().map(|(name, e)| (name, e.0)).collect();
+        self.inner = self.inner.clone().select(Select::Expr(pairs));
+    }
+
     pub fn select_columns(&mut self, columns: Vec<String>) {
         self.inner = self.inner.clone().select(Select::columns(&columns));
     }
@@ -712,6 +747,10 @@ impl VectorQuery {
         self.inner = self.inner.clone().only_if(predicate);
     }
 
+    pub fn where_expr(&mut self, expr: PyExpr) {
+        self.inner = self.inner.clone().only_if_expr(expr.0);
+    }
+
     pub fn add_query_vector(&mut self, vector: Bound<'_, PyAny>) -> PyResult<()> {
         let data: ArrayData = ArrayData::from_pyarrow_bound(&vector)?;
         let array = make_array(data);
@@ -723,6 +762,12 @@ impl VectorQuery {
         self.inner = self.inner.clone().select(Select::dynamic(&columns));
     }
 
+    pub fn select_expr(&mut self, columns: Vec<(String, PyExpr)>) {
+        let pairs: Vec<(String, lancedb::expr::DfExpr)> =
+            columns.into_iter().map(|(name, e)| (name, e.0)).collect();
+        self.inner = self.inner.clone().select(Select::Expr(pairs));
+    }
+
     pub fn select_columns(&mut self, columns: Vec<String>) {
         self.inner = self.inner.clone().select(Select::columns(&columns));
     }
@@ -877,11 +922,21 @@ impl HybridQuery {
         self.inner_fts.r#where(predicate);
     }
 
+    pub fn where_expr(&mut self, expr: PyExpr) {
+        self.inner_vec.where_expr(expr.clone());
+        self.inner_fts.where_expr(expr);
+    }
+
     pub fn select(&mut self, columns: Vec<(String, String)>) {
         self.inner_vec.select(columns.clone());
         self.inner_fts.select(columns);
     }
 
+    pub fn select_expr(&mut self, columns: Vec<(String, PyExpr)>) {
+        self.inner_vec.select_expr(columns.clone());
+        self.inner_fts.select_expr(columns);
+    }
+
     pub fn select_columns(&mut self, columns: Vec<String>) {
         self.inner_vec.select_columns(columns.clone());
         self.inner_fts.select_columns(columns);

python/src/session.rs

@@ -4,7 +4,7 @@
 use std::sync::Arc;
 
 use lancedb::{ObjectStoreRegistry, Session as LanceSession};
-use pyo3::{pyclass, pymethods, PyResult};
+use pyo3::{PyResult, pyclass, pymethods};
 
 /// A session for managing caches and object stores across LanceDB operations.
 ///

python/src/storage_options.rs

@@ -1,150 +0,0 @@
-// SPDX-License-Identifier: Apache-2.0
-// SPDX-FileCopyrightText: Copyright The LanceDB Authors
-
-//! PyO3 bindings for StorageOptionsProvider
-//!
-//! This module provides the bridge between Python StorageOptionsProvider objects
-//! and Rust's StorageOptionsProvider trait, enabling automatic credential refresh.
-
-use std::collections::HashMap;
-use std::sync::Arc;
-
-use async_trait::async_trait;
-use lance_io::object_store::StorageOptionsProvider;
-use pyo3::prelude::*;
-use pyo3::types::PyDict;
-
-/// Internal wrapper around a Python object implementing StorageOptionsProvider
-pub struct PyStorageOptionsProvider {
-    /// The Python object implementing fetch_storage_options()
-    inner: Py<PyAny>,
-}
-
-impl Clone for PyStorageOptionsProvider {
-    fn clone(&self) -> Self {
-        Python::attach(|py| Self {
-            inner: self.inner.clone_ref(py),
-        })
-    }
-}
-
-impl PyStorageOptionsProvider {
-    pub fn new(obj: Py<PyAny>) -> PyResult<Self> {
-        Python::attach(|py| {
-            // Verify the object has a fetch_storage_options method
-            if !obj.bind(py).hasattr("fetch_storage_options")? {
-                return Err(pyo3::exceptions::PyTypeError::new_err(
-                    "StorageOptionsProvider must implement fetch_storage_options() method",
-                ));
-            }
-            Ok(Self { inner: obj })
-        })
-    }
-}
-
-/// Wrapper that implements the Rust StorageOptionsProvider trait
-pub struct PyStorageOptionsProviderWrapper {
-    py_provider: PyStorageOptionsProvider,
-}
-
-impl PyStorageOptionsProviderWrapper {
-    pub fn new(py_provider: PyStorageOptionsProvider) -> Self {
-        Self { py_provider }
-    }
-}
-
-#[async_trait]
-impl StorageOptionsProvider for PyStorageOptionsProviderWrapper {
-    async fn fetch_storage_options(&self) -> lance_core::Result<Option<HashMap<String, String>>> {
-        // Call Python method from async context using spawn_blocking
-        let py_provider = self.py_provider.clone();
-
-        tokio::task::spawn_blocking(move || {
-            Python::attach(|py| {
-                // Call the Python fetch_storage_options method
-                let result = py_provider
-                    .inner
-                    .bind(py)
-                    .call_method0("fetch_storage_options")
-                    .map_err(|e| lance_core::Error::IO {
-                        source: Box::new(std::io::Error::other(format!(
-                            "Failed to call fetch_storage_options: {}",
-                            e
-                        ))),
-                        location: snafu::location!(),
-                    })?;
-
-                // If result is None, return None
-                if result.is_none() {
-                    return Ok(None);
-                }
-
-                // Extract the result dict - should be a flat Map<String, String>
-                let result_dict = result.downcast::<PyDict>().map_err(|_| {
-                    lance_core::Error::InvalidInput {
-                        source: "fetch_storage_options() must return None or a dict of string key-value pairs".into(),
-                        location: snafu::location!(),
-                    }
-                })?;
-
-                // Convert all entries to HashMap<String, String>
-                let mut storage_options = HashMap::new();
-                for (key, value) in result_dict.iter() {
-                    let key_str: String = key.extract().map_err(|e| {
-                        lance_core::Error::InvalidInput {
-                            source: format!("Storage option key must be a string: {}", e).into(),
-                            location: snafu::location!(),
-                        }
-                    })?;
-                    let value_str: String = value.extract().map_err(|e| {
-                        lance_core::Error::InvalidInput {
-                            source: format!("Storage option value must be a string: {}", e).into(),
-                            location: snafu::location!(),
-                        }
-                    })?;
-                    storage_options.insert(key_str, value_str);
-                }
-
-                Ok(Some(storage_options))
-            })
-        })
-        .await
-        .map_err(|e| lance_core::Error::IO {
-            source: Box::new(std::io::Error::other(format!(
-                "Task join error: {}",
-                e
-            ))),
-            location: snafu::location!(),
-        })?
-    }
-
-    fn provider_id(&self) -> String {
-        Python::attach(|py| {
-            // Call provider_id() method on the Python object
-            let obj = self.py_provider.inner.bind(py);
-            obj.call_method0("provider_id")
-                .and_then(|result| result.extract::<String>())
-                .unwrap_or_else(|e| {
-                    // If provider_id() fails, construct a fallback ID
-                    format!("PyStorageOptionsProvider(error: {})", e)
-                })
-        })
-    }
-}
-
-impl std::fmt::Debug for PyStorageOptionsProviderWrapper {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "PyStorageOptionsProviderWrapper({})", self.provider_id())
-    }
-}
-
-/// Convert a Python object to an Arc<dyn StorageOptionsProvider>
-///
-/// This is the main entry point for converting Python StorageOptionsProvider objects
-/// to Rust trait objects that can be used by the Lance ecosystem.
-pub fn py_object_to_storage_options_provider(
-    py_obj: Py<PyAny>,
-) -> PyResult<Arc<dyn StorageOptionsProvider>> {
-    let py_provider = PyStorageOptionsProvider::new(py_obj)?;
-    Ok(Arc::new(PyStorageOptionsProviderWrapper::new(py_provider)))
-}

python/src/table.rs

@@ -5,8 +5,9 @@ use std::{collections::HashMap, sync::Arc};
 use crate::{
     connection::Connection,
     error::PythonErrorExt,
-    index::{extract_index_params, IndexConfig},
+    index::{IndexConfig, extract_index_params},
     query::{Query, TakeQuery},
+    table::scannable::PyScannable,
 };
 use arrow::{
     datatypes::{DataType, Schema},
@@ -18,13 +19,15 @@ use lancedb::table::{
     Table as LanceDbTable,
 };
 use pyo3::{
+    Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
     exceptions::{PyKeyError, PyRuntimeError, PyValueError},
     pyclass, pymethods,
     types::{IntoPyDict, PyAnyMethods, PyDict, PyDictMethods},
-    Bound, FromPyObject, PyAny, PyRef, PyResult, Python,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 
+mod scannable;
+
 /// Statistics about a compaction operation.
 #[pyclass(get_all)]
 #[derive(Clone, Debug)]
@@ -109,19 +112,24 @@ impl From<lancedb::table::AddResult> for AddResult {
 #[pyclass(get_all)]
 #[derive(Clone, Debug)]
 pub struct DeleteResult {
+    pub num_deleted_rows: u64,
     pub version: u64,
 }
 
 #[pymethods]
 impl DeleteResult {
     pub fn __repr__(&self) -> String {
-        format!("DeleteResult(version={})", self.version)
+        format!(
+            "DeleteResult(num_deleted_rows={}, version={})",
+            self.num_deleted_rows, self.version
+        )
     }
 }
 
 impl From<lancedb::table::DeleteResult> for DeleteResult {
     fn from(result: lancedb::table::DeleteResult) -> Self {
         Self {
+            num_deleted_rows: result.num_deleted_rows,
             version: result.version,
         }
     }
@@ -291,14 +299,14 @@ impl Table {
         })
     }
 
+    #[pyo3(signature = (data, mode, progress=None))]
     pub fn add<'a>(
         self_: PyRef<'a, Self>,
-        data: Bound<'_, PyAny>,
+        data: PyScannable,
         mode: String,
+        progress: Option<Py<PyAny>>,
     ) -> PyResult<Bound<'a, PyAny>> {
-        let batches: Box<dyn arrow::array::RecordBatchReader + Send> =
+        let mut op = self_.inner_ref()?.add(data);
-            Box::new(ArrowArrayStreamReader::from_pyarrow_bound(&data)?);
-        let mut op = self_.inner_ref()?.add(batches);
         if mode == "append" {
             op = op.mode(AddDataMode::Append);
         } else if mode == "overwrite" {
@@ -306,6 +314,81 @@ impl Table {
         } else {
             return Err(PyValueError::new_err(format!("Invalid mode: {}", mode)));
         }
+        if let Some(progress_obj) = progress {
+            let is_callable = Python::attach(|py| progress_obj.bind(py).is_callable());
+            if is_callable {
+                // Callback: call with a dict of progress info.
+                op = op.progress(move |p| {
+                    Python::attach(|py| {
+                        let dict = PyDict::new(py);
+                        if let Err(e) = dict
+                            .set_item("output_rows", p.output_rows())
+                            .and_then(|_| dict.set_item("output_bytes", p.output_bytes()))
+                            .and_then(|_| dict.set_item("total_rows", p.total_rows()))
+                            .and_then(|_| {
+                                dict.set_item("elapsed_seconds", p.elapsed().as_secs_f64())
+                            })
+                            .and_then(|_| dict.set_item("active_tasks", p.active_tasks()))
+                            .and_then(|_| dict.set_item("total_tasks", p.total_tasks()))
+                            .and_then(|_| dict.set_item("done", p.done()))
+                        {
+                            log::warn!("progress dict error: {e}");
+                            return;
+                        }
+                        if let Err(e) = progress_obj.call1(py, (dict,)) {
+                            log::warn!("progress callback error: {e}");
+                        }
+                    });
+                });
+            } else {
+                // tqdm-like: has update() method.
+                let mut last_rows: usize = 0;
+                let mut total_set = false;
+                op = op.progress(move |p| {
+                    let current = p.output_rows();
+                    let prev = last_rows;
+                    last_rows = current;
+                    Python::attach(|py| {
+                        if let Some(total) = p.total_rows()
+                            && !total_set
+                        {
+                            if let Err(e) = progress_obj.setattr(py, "total", total) {
+                                log::warn!("progress setattr error: {e}");
+                            }
+                            total_set = true;
+                        }
+                        let delta = current.saturating_sub(prev);
+                        if delta > 0 {
+                            if let Err(e) = progress_obj.call_method1(py, "update", (delta,)) {
+                                log::warn!("progress update error: {e}");
+                            }
+                            // Show throughput and active workers in tqdm postfix.
+                            let elapsed = p.elapsed().as_secs_f64();
+                            if elapsed > 0.0 {
+                                let mb_per_sec = p.output_bytes() as f64 / elapsed / 1_000_000.0;
+                                let postfix = format!(
+                                    "{:.1} MB/s | {}/{} workers",
+                                    mb_per_sec,
+                                    p.active_tasks(),
+                                    p.total_tasks()
+                                );
+                                if let Err(e) =
+                                    progress_obj.call_method1(py, "set_postfix_str", (postfix,))
+                                {
+                                    log::warn!("progress set_postfix_str error: {e}");
+                                }
+                            }
+                        }
+                        if p.done() {
+                            // Force a final refresh so the bar shows completion.
+                            if let Err(e) = progress_obj.call_method0(py, "refresh") {
+                                log::warn!("progress refresh error: {e}");
+                            }
+                        }
+                    });
+                });
+            }
+        }
 
         future_into_py(self_.py(), async move {
             let result = op.execute().await.infer_error()?;
@@ -420,6 +503,17 @@ impl Table {
         })
     }
 
+    pub fn prewarm_data(
+        self_: PyRef<'_, Self>,
+        columns: Option<Vec<String>>,
+    ) -> PyResult<Bound<'_, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        future_into_py(self_.py(), async move {
+            inner.prewarm_data(columns).await.infer_error()?;
+            Ok(())
+        })
+    }
+
     pub fn list_indices(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
@@ -536,7 +630,7 @@ impl Table {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
             let versions = inner.list_versions().await.infer_error()?;
-            let versions_as_dict = Python::attach(|py| {
+            Python::attach(|py| {
                 versions
                     .iter()
                     .map(|v| {
@@ -553,9 +647,7 @@ impl Table {
                         Ok(dict.unbind())
                     })
                     .collect::<PyResult<Vec<_>>>()
-            });
+            })
-
-            versions_as_dict
         })
     }
 

python/src/table/scannable.rs

@@ -0,0 +1,145 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+use std::sync::Arc;
+
+use arrow::{
+    datatypes::{Schema, SchemaRef},
+    ffi_stream::ArrowArrayStreamReader,
+    pyarrow::{FromPyArrow, PyArrowType},
+};
+use futures::StreamExt;
+use lancedb::{
+    Error,
+    arrow::{SendableRecordBatchStream, SimpleRecordBatchStream},
+    data::scannable::Scannable,
+};
+use pyo3::{FromPyObject, Py, PyAny, Python, types::PyAnyMethods};
+
+/// Adapter that implements Scannable for a Python reader factory callable.
+///
+/// This holds a Python callable that returns a RecordBatchReader when called.
+/// For rescannable sources, the callable can be invoked multiple times to
+/// get fresh readers.
+pub struct PyScannable {
+    /// Python callable that returns a RecordBatchReader
+    reader_factory: Py<PyAny>,
+    schema: SchemaRef,
+    num_rows: Option<usize>,
+    rescannable: bool,
+}
+
+impl std::fmt::Debug for PyScannable {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("PyScannable")
+            .field("schema", &self.schema)
+            .field("num_rows", &self.num_rows)
+            .field("rescannable", &self.rescannable)
+            .finish()
+    }
+}
+
+impl Scannable for PyScannable {
+    fn schema(&self) -> SchemaRef {
+        self.schema.clone()
+    }
+
+    fn scan_as_stream(&mut self) -> SendableRecordBatchStream {
+        let reader: Result<ArrowArrayStreamReader, Error> = {
+            Python::attach(|py| {
+                let result =
+                    self.reader_factory
+                        .call0(py)
+                        .map_err(|e| lancedb::Error::Runtime {
+                            message: format!("Python reader factory failed: {}", e),
+                        })?;
+                ArrowArrayStreamReader::from_pyarrow_bound(result.bind(py)).map_err(|e| {
+                    lancedb::Error::Runtime {
+                        message: format!("Failed to create Arrow reader from Python: {}", e),
+                    }
+                })
+            })
+        };
+
+        // Reader is blocking but stream is non-blocking, so we need to spawn a task to pull.
+        let (tx, rx) = tokio::sync::mpsc::channel(1);
+
+        let join_handle = tokio::task::spawn_blocking(move || {
+            let reader = match reader {
+                Ok(reader) => reader,
+                Err(e) => {
+                    let _ = tx.blocking_send(Err(e));
+                    return;
+                }
+            };
+            for batch in reader {
+                match batch {
+                    Ok(batch) => {
+                        if tx.blocking_send(Ok(batch)).is_err() {
+                            // Receiver dropped, stop processing
+                            break;
+                        }
+                    }
+                    Err(source) => {
+                        let _ = tx.blocking_send(Err(Error::Arrow { source }));
+                        break;
+                    }
+                }
+            }
+        });
+
+        let schema = self.schema.clone();
+        let stream = futures::stream::unfold(
+            (rx, Some(join_handle)),
+            |(mut rx, join_handle)| async move {
+                match rx.recv().await {
+                    Some(Ok(batch)) => Some((Ok(batch), (rx, join_handle))),
+                    Some(Err(e)) => Some((Err(e), (rx, join_handle))),
+                    None => {
+                        // Channel closed. Check if the task panicked — a panic
+                        // drops the sender without sending an error, so without
+                        // this check we'd silently return a truncated stream.
+                        if let Some(handle) = join_handle
+                            && let Err(join_err) = handle.await
+                        {
+                            return Some((
+                                Err(Error::Runtime {
+                                    message: format!("Reader task panicked: {}", join_err),
+                                }),
+                                (rx, None),
+                            ));
+                        }
+                        None
+                    }
+                }
+            },
+        );
+        Box::pin(SimpleRecordBatchStream::new(stream.fuse(), schema))
+    }
+
+    fn num_rows(&self) -> Option<usize> {
+        self.num_rows
+    }
+
+    fn rescannable(&self) -> bool {
+        self.rescannable
+    }
+}
+
+impl<'py> FromPyObject<'py> for PyScannable {
+    fn extract_bound(ob: &pyo3::Bound<'py, PyAny>) -> pyo3::PyResult<Self> {
+        // Convert from Scannable dataclass.
+        let schema: PyArrowType<Schema> = ob.getattr("schema")?.extract()?;
+        let schema = Arc::new(schema.0);
+        let num_rows: Option<usize> = ob.getattr("num_rows")?.extract()?;
+        let rescannable: bool = ob.getattr("rescannable")?.extract()?;
+        let reader_factory: Py<PyAny> = ob.getattr("reader")?.unbind();
+
+        Ok(Self {
+            schema,
+            reader_factory,
+            num_rows,
+            rescannable,
+        })
+    }
+}