Bump version: 0.29.2-beta.0 → 0.29.2

## Summary

- Added `repository` field to all nodejs package.json files (main
package + 7 platform-specific packages)
- This fixes the npm publish E422 error where sigstore provenance
verification fails because the repository.url was empty

## Root Cause

Failing CI:
https://github.com/lancedb/lancedb/actions/runs/21770794768/job/62821570260

npm's sigstore provenance verification requires the `repository.url`
field in package.json to match the GitHub repository URL from the
provenance bundle. The platform-specific packages
(`@lancedb/lancedb-darwin-arm64`, etc.) were missing this field
entirely, causing the publish to fail with:

```
npm error 422 Unprocessable Entity - Error verifying sigstore provenance bundle: 
Failed to validate repository information: package.json: "repository.url" is "", 
expected to match "https://github.com/lancedb/lancedb" from provenance
```

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.23.1"
+current_version = "0.26.1"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/workflows/build_linux_wheel/action.yml

@@ -3,7 +3,7 @@ name: build-linux-wheel
 description: "Build a manylinux wheel for lance"
 inputs:
   python-minor-version:
-    description: "8, 9, 10, 11, 12"
+    description: "10, 11, 12, 13"
     required: true
   args:
     description: "--release"

.github/workflows/build_mac_wheel/action.yml

@@ -3,7 +3,7 @@ name: build_wheel
 description: "Build a lance wheel"
 inputs:
   python-minor-version:
-    description: "8, 9, 10, 11"
+    description: "10, 11, 12, 13"
     required: true
   args:
     description: "--release"

.github/workflows/build_windows_wheel/action.yml

@@ -3,7 +3,7 @@ name: build_wheel
 description: "Build a lance wheel"
 inputs:
   python-minor-version:
-    description: "8, 9, 10, 11"
+    description: "10, 11, 12, 13, 14"
     required: true
   args:
     description: "--release"

.github/workflows/cargo-publish.yml

@@ -42,7 +42,7 @@ jobs:
     name: Report Workflow Failure
     runs-on: ubuntu-latest
     needs: [build]
-    if: always() && (github.event_name == 'release' || github.event_name == 'workflow_dispatch')
+    if: always() && failure() && startsWith(github.ref, 'refs/tags/v')
     permissions:
       contents: read
       issues: write

.github/workflows/codex-update-lance-dependency.yml

@@ -75,20 +75,28 @@ jobs:
           VERSION="${VERSION#v}"
           BRANCH_NAME="codex/update-lance-${VERSION//[^a-zA-Z0-9]/-}"
 
+          # Use "chore" for beta/rc versions, "feat" for stable releases
+          if [[ "${VERSION}" == *beta* ]] || [[ "${VERSION}" == *rc* ]]; then
+            COMMIT_TYPE="chore"
+          else
+            COMMIT_TYPE="feat"
+          fi
+
           cat <<EOF >/tmp/codex-prompt.txt
           You are running inside the lancedb repository on a GitHub Actions runner. Update the Lance dependency to version ${VERSION} and prepare a pull request for maintainers to review.
 
           Follow these steps exactly:
-          1. Use script "ci/set_lance_version.py" to update Lance dependencies. The script already refreshes Cargo metadata, so allow it to finish even if it takes time.
-          2. Run "cargo clippy --workspace --tests --all-features -- -D warnings". If diagnostics appear, fix them yourself and rerun clippy until it exits cleanly. Do not skip any warnings.
-          3. After clippy succeeds, run "cargo fmt --all" to format the workspace.
-          4. Ensure the repository is clean except for intentional changes. Inspect "git status --short" and "git diff" to confirm the dependency update and any required fixes.
-          5. Create and switch to a new branch named "${BRANCH_NAME}" (replace any duplicated hyphens if necessary).
-          6. Stage all relevant files with "git add -A". Commit using the message "chore: update lance dependency to v${VERSION}".
-          7. Push the branch to origin. If the branch already exists, force-push your changes.
-          8. env "GH_TOKEN" is available, use "gh" tools for github related operations like creating pull request.
-          9. Create a pull request targeting "main" with title "chore: update lance dependency to v${VERSION}". In the body, summarize the dependency bump, clippy/fmt verification, and link the triggering tag (${TAG}).
-          10. After creating the PR, display the PR URL, "git status --short", and a concise summary of the commands run and their results.
+          1. Use script "ci/set_lance_version.py" to update Lance Rust dependencies. The script already refreshes Cargo metadata, so allow it to finish even if it takes time.
+          2. Update the Java lance-core dependency version in "java/pom.xml": change the "<lance-core.version>...</lance-core.version>" property to "${VERSION}".
+          3. Run "cargo clippy --workspace --tests --all-features -- -D warnings". If diagnostics appear, fix them yourself and rerun clippy until it exits cleanly. Do not skip any warnings.
+          4. After clippy succeeds, run "cargo fmt --all" to format the workspace.
+          5. Ensure the repository is clean except for intentional changes. Inspect "git status --short" and "git diff" to confirm the dependency update and any required fixes.
+          6. Create and switch to a new branch named "${BRANCH_NAME}" (replace any duplicated hyphens if necessary).
+          7. Stage all relevant files with "git add -A". Commit using the message "${COMMIT_TYPE}: update lance dependency to v${VERSION}".
+          8. Push the branch to origin. If the remote branch already exists, delete it first with "gh api -X DELETE repos/lancedb/lancedb/git/refs/heads/${BRANCH_NAME}" then push with "git push origin ${BRANCH_NAME}". Do NOT use "git push --force" or "git push -f".
+          9. env "GH_TOKEN" is available, use "gh" tools for github related operations like creating pull request.
+          10. Create a pull request targeting "main" with title "${COMMIT_TYPE}: update lance dependency to v${VERSION}". First, write the PR body to /tmp/pr-body.md using a heredoc (cat <<'EOF' > /tmp/pr-body.md). The body should summarize the dependency bump, clippy/fmt verification, and link the triggering tag (${TAG}). Then run "gh pr create --body-file /tmp/pr-body.md".
+          11. After creating the PR, display the PR URL, "git status --short", and a concise summary of the commands run and their results.
 
           Constraints:
           - Use bash commands; avoid modifying GitHub workflow files other than through the scripted task above.

.github/workflows/docs.yml

@@ -41,7 +41,7 @@ jobs:
           sudo apt install -y protobuf-compiler libssl-dev
           rustup update && rustup default
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: "3.10"
           cache: "pip"

.github/workflows/nodejs.yml

@@ -8,6 +8,7 @@ on:
     paths:
       - Cargo.toml
       - nodejs/**
+      - docs/src/js/**
       - .github/workflows/nodejs.yml
       - docker-compose.yml
 

.github/workflows/npm-publish.yml

@@ -318,7 +318,7 @@ jobs:
       - name: Setup node
         uses: actions/setup-node@v4
         with:
-          node-version: 20
+          node-version: 24
           cache: npm
           cache-dependency-path: nodejs/package-lock.json
           registry-url: "https://registry.npmjs.org"
@@ -348,9 +348,9 @@ jobs:
         run: find npm
       - name: Publish
         env:
-          NODE_AUTH_TOKEN: ${{ secrets.LANCEDB_NPM_REGISTRY_TOKEN }}
           DRY_RUN: ${{ !startsWith(github.ref, 'refs/tags/v') }}
         run: |
+          npm config set provenance true
           ARGS="--access public"
           if [[ $DRY_RUN == "true" ]]; then
             ARGS="$ARGS --dry-run"
@@ -363,7 +363,7 @@ jobs:
     name: Report Workflow Failure
     runs-on: ubuntu-latest
     needs: [build-lancedb, test-lancedb, publish]
-    if: always() && (github.event_name == 'release' || github.event_name == 'workflow_dispatch')
+    if: always() && failure() && startsWith(github.ref, 'refs/tags/v')
     permissions:
       contents: read
       issues: write

.github/workflows/pypi-publish.yml

@@ -44,12 +44,12 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v6
         with:
-          python-version: 3.8
+          python-version: "3.10"
       - uses: ./.github/workflows/build_linux_wheel
         with:
-          python-minor-version: 8
+          python-minor-version: 10
           args: "--release --strip ${{ matrix.config.extra_args }}"
           arm-build: ${{ matrix.config.platform == 'aarch64' }}
           manylinux: ${{ matrix.config.manylinux }}
@@ -74,12 +74,12 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v6
         with:
-          python-version: 3.12
+          python-version: "3.13"
       - uses: ./.github/workflows/build_mac_wheel
         with:
-          python-minor-version: 8
+          python-minor-version: 10
           args: "--release --strip --target ${{ matrix.config.target }} --features fp16kernels"
       - uses: ./.github/workflows/upload_wheel
         if: startsWith(github.ref, 'refs/tags/python-v')
@@ -95,12 +95,12 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v6
         with:
-          python-version: 3.12
+          python-version: "3.13"
       - uses: ./.github/workflows/build_windows_wheel
         with:
-          python-minor-version: 8
+          python-minor-version: 10
           args: "--release --strip"
           vcpkg_token: ${{ secrets.VCPKG_GITHUB_PACKAGES }}
       - uses: ./.github/workflows/upload_wheel
@@ -181,7 +181,7 @@ jobs:
     permissions:
       contents: read
       issues: write
-    if: always() && (github.event_name == 'release' || github.event_name == 'workflow_dispatch')
+    if: always() && failure() && startsWith(github.ref, 'refs/tags/python-v')
     steps:
       - uses: actions/checkout@v4
       - uses: ./.github/actions/create-failure-issue

.github/workflows/python.yml

@@ -36,9 +36,9 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: "3.12"
+          python-version: "3.13"
       - name: Install ruff
         run: |
           pip install ruff==0.9.9
@@ -61,9 +61,9 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: "3.12"
+          python-version: "3.13"
       - name: Install protobuf compiler
         run: |
           sudo apt update
@@ -90,9 +90,9 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: "3.12"
+          python-version: "3.13"
           cache: "pip"
       - name: Install protobuf
         run: |
@@ -110,7 +110,7 @@ jobs:
     timeout-minutes: 30
     strategy:
       matrix:
-        python-minor-version: ["9", "12"]
+        python-minor-version: ["10", "13"]
     runs-on: "ubuntu-24.04"
     defaults:
       run:
@@ -126,7 +126,7 @@ jobs:
           sudo apt update
           sudo apt install -y protobuf-compiler
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: 3.${{ matrix.python-minor-version }}
       - uses: ./.github/workflows/build_linux_wheel
@@ -156,9 +156,9 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: "3.12"
+          python-version: "3.13"
       - uses: ./.github/workflows/build_mac_wheel
         with:
           args: --profile ci
@@ -185,9 +185,9 @@ jobs:
           fetch-depth: 0
           lfs: true
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: "3.12"
+          python-version: "3.13"
       - uses: ./.github/workflows/build_windows_wheel
         with:
           args: --profile ci
@@ -212,9 +212,9 @@ jobs:
           sudo apt update
           sudo apt install -y protobuf-compiler
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
-          python-version: 3.9
+          python-version: "3.10"
       - name: Install lancedb
         run: |
           pip install "pydantic<2"

.github/workflows/rust.yml

@@ -48,6 +48,8 @@ jobs:
         run: cargo fmt --all -- --check
       - name: Run clippy
         run: cargo clippy --profile ci --workspace --tests --all-features -- -D warnings
+      - name: Run clippy (without remote feature)
+        run: cargo clippy --profile ci --workspace --tests -- -D warnings
 
   build-no-lock:
     runs-on: ubuntu-24.04
@@ -181,7 +183,7 @@ jobs:
     runs-on: ubuntu-24.04
     strategy:
       matrix:
-        msrv: ["1.78.0"] # This should match up with rust-version in Cargo.toml
+        msrv: ["1.88.0"] # This should match up with rust-version in Cargo.toml
     env:
       # Need up-to-date compilers for kernels
       CC: clang-18
@@ -212,4 +214,6 @@ jobs:
           cargo update -p aws-sdk-sts --precise 1.51.0
           cargo update -p home --precise 0.5.9
       - name: cargo +${{ matrix.msrv }} check
+        env:
+          RUSTUP_TOOLCHAIN: ${{ matrix.msrv }}
         run: cargo check --profile ci --workspace --tests --benches --all-features

Cargo.lock

@@ -1,6 +1,6 @@
 # This file is automatically @generated by Cargo.
 # It is not intended for manual editing.
-version = 3
+version = 4
 
 [[package]]
 name = "adler2"
@@ -3072,8 +3072,9 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
 
 [[package]]
 name = "fsst"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0f03a771ab914e207dd26bd2f12666839555ec8ecc7e1770e1ed6f9900d899a4"
 dependencies = [
  "arrow-array",
  "rand 0.9.2",
@@ -4404,8 +4405,9 @@ dependencies = [
 
 [[package]]
 name = "lance"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "47b685aca3f97ee02997c83ded16f59c747ccb69e74c8abbbae4aa3d22cf1301"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4470,8 +4472,9 @@ dependencies = [
 
 [[package]]
 name = "lance-arrow"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "daf00c7537df524cc518a089f0d156a036d95ca3f5bc2bc1f0a9f9293e9b62ef"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4490,8 +4493,9 @@ dependencies = [
 
 [[package]]
 name = "lance-bitpacking"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "46752e4ac8fc5590a445e780b63a8800adc7a770bd74770a8dc66963778e4e77"
 dependencies = [
  "arrayref",
  "paste",
@@ -4500,8 +4504,9 @@ dependencies = [
 
 [[package]]
 name = "lance-core"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3d13d87d07305c6d4b4dc7780fb1107babf782a0e5b1dc7872e17ae1f8fd11ca"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4538,8 +4543,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datafusion"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6451b5af876eaef8bec4b38a39dadac9d44621e1ecf85d0cdf6097a5d0aa8721"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4569,8 +4575,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datagen"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e1736708dd7867dfbab8fcc930b21c96717c6c00be73b7d9a240336a4ed80375"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4588,8 +4595,9 @@ dependencies = [
 
 [[package]]
 name = "lance-encoding"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d6b6ca4ff94833240d5ba4a94a742cba786d1949b3c3fa7e11d6f0050443432a"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4626,8 +4634,9 @@ dependencies = [
 
 [[package]]
 name = "lance-file"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "55fbe959bffe185543aed3cbeb14484f1aa2e55886034fdb1ea3d8cc9b70aad8"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4659,8 +4668,9 @@ dependencies = [
 
 [[package]]
 name = "lance-geo"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a52b0adabc953d457f336a784a3b37353a180e6a79905f544949746e0d4c6483"
 dependencies = [
  "datafusion",
  "geo-traits",
@@ -4674,8 +4684,9 @@ dependencies = [
 
 [[package]]
 name = "lance-index"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6b67654bf86fd942dd2cf08294ee7e91053427cd148225f49c9ff398ff9a40fd"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4742,8 +4753,9 @@ dependencies = [
 
 [[package]]
 name = "lance-io"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8eb0ccc1c414e31687d83992d546af0a0237c8d2f4bf2ae3d347d539fd0fc141"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4783,8 +4795,9 @@ dependencies = [
 
 [[package]]
 name = "lance-linalg"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "083404cf12dcdb1a7df98fb58f9daf626b6e43a2f794b37b6b89b4012a0e1f78"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4800,8 +4813,9 @@ dependencies = [
 
 [[package]]
 name = "lance-namespace"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c12778d2aabf9c2bfd16e2509ebe120e562a288d8ae630ec6b6b204868df41b2"
 dependencies = [
  "arrow",
  "async-trait",
@@ -4813,8 +4827,9 @@ dependencies = [
 
 [[package]]
 name = "lance-namespace-impls"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8863aababdd13a6d2c8d6179dc6981f4f8f49d8b66a00c5dd75115aec4cadc99"
 dependencies = [
  "arrow",
  "arrow-ipc",
@@ -4857,8 +4872,9 @@ dependencies = [
 
 [[package]]
 name = "lance-table"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f0fcc83f197ce2000c4abe4f5e0873490ab1f41788fa76571c4209b87d4daf50"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4897,8 +4913,9 @@ dependencies = [
 
 [[package]]
 name = "lance-testing"
-version = "2.0.0-beta.8"
-source = "git+https://github.com/lance-format/lance.git?tag=v2.0.0-beta.8#5bffcfc3f9716fd6eb069579eb5027ed9c69feb7"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7fb1f7c7e06f91360e141ecee1cf2110f858c231705f69f2cd2fda9e30c1e9f4"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -4909,7 +4926,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb"
-version = "0.23.1"
+version = "0.26.1"
 dependencies = [
  "ahash",
  "anyhow",
@@ -4939,6 +4956,7 @@ dependencies = [
  "datafusion-common",
  "datafusion-execution",
  "datafusion-expr",
+ "datafusion-physical-expr",
  "datafusion-physical-plan",
  "futures",
  "half",
@@ -4988,7 +5006,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb-nodejs"
-version = "0.23.1"
+version = "0.26.1"
 dependencies = [
  "arrow-array",
  "arrow-ipc",
@@ -5008,7 +5026,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb-python"
-version = "0.26.1"
+version = "0.29.1"
 dependencies = [
  "arrow",
  "async-trait",

Cargo.toml

@@ -12,23 +12,23 @@ repository = "https://github.com/lancedb/lancedb"
 description = "Serverless, low-latency vector database for AI applications"
 keywords = ["lancedb", "lance", "database", "vector", "search"]
 categories = ["database-implementations"]
-rust-version = "1.78.0"
+rust-version = "1.88.0"
 
 [workspace.dependencies]
-lance = { "version" = "=2.0.0-beta.8", default-features = false, "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=2.0.0-beta.8", default-features = false, "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=2.0.0-beta.8", default-features = false, "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=2.0.0-beta.8", "tag" = "v2.0.0-beta.8", "git" = "https://github.com/lance-format/lance.git" }
+lance = { "version" = "=2.0.0", default-features = false }
+lance-core = "=2.0.0"
+lance-datagen = "=2.0.0"
+lance-file = "=2.0.0"
+lance-io = { "version" = "=2.0.0", default-features = false }
+lance-index = "=2.0.0"
+lance-linalg = "=2.0.0"
+lance-namespace = "=2.0.0"
+lance-namespace-impls = { "version" = "=2.0.0", default-features = false }
+lance-table = "=2.0.0"
+lance-testing = "=2.0.0"
+lance-datafusion = "=2.0.0"
+lance-encoding = "=2.0.0"
+lance-arrow = "=2.0.0"
 ahash = "0.8"
 # Note that this one does not include pyarrow
 arrow = { version = "57.2", optional = false }
@@ -46,6 +46,7 @@ datafusion-common = { version = "51.0", default-features = false }
 datafusion-execution = "51.0"
 datafusion-expr = "51.0"
 datafusion-physical-plan = "51.0"
+datafusion-physical-expr = "51.0"
 env_logger = "0.11"
 half = { "version" = "2.7.1", default-features = false, features = [
     "num-traits",
@@ -59,7 +60,7 @@ rand = "0.9"
 snafu = "0.8"
 url = "2"
 num-traits = "0.2"
-regex = "1.12"
+regex = "1.10"
 lazy_static = "1"
 semver = "1.0.25"
 chrono = "0.4"

README.md

@@ -66,7 +66,7 @@ Follow the [Quickstart](https://lancedb.com/docs/quickstart/) doc to set up Lanc
 | Python SDK | https://lancedb.github.io/lancedb/python/python/ |
 | Typescript SDK | https://lancedb.github.io/lancedb/js/globals/ |
 | Rust SDK | https://docs.rs/lancedb/latest/lancedb/index.html |
-| REST API | https://docs.lancedb.com/api-reference/introduction |
+| REST API | https://docs.lancedb.com/api-reference/rest |
 
 ## **Join Us and Contribute**
 

docs/src/embeddings/available_embedding_models/text_embedding_functions/voyageai_embedding.md

@@ -0,0 +1,62 @@
+# VoyageAI Embeddings
+
+Voyage AI provides cutting-edge embedding and rerankers.
+
+
+Using voyageai API requires voyageai package, which can be installed using `pip install voyageai`. Voyage AI embeddings are used to generate embeddings for text data. The embeddings can be used for various tasks like semantic search, clustering, and classification.
+You also need to set the `VOYAGE_API_KEY` environment variable to use the VoyageAI API.
+
+Supported models are:
+
+**Voyage-4 Series (Latest)**
+
+- voyage-4 (1024 dims, general-purpose and multilingual retrieval, 320K batch tokens)
+- voyage-4-lite (1024 dims, optimized for latency and cost, 1M batch tokens)
+- voyage-4-large (1024 dims, best retrieval quality, 120K batch tokens)
+
+**Voyage-3 Series**
+
+- voyage-3
+- voyage-3-lite
+
+**Domain-Specific Models**
+
+- voyage-finance-2
+- voyage-multilingual-2
+- voyage-law-2
+- voyage-code-2
+
+
+Supported parameters (to be passed in `create` method) are:
+
+| Parameter | Type | Default Value | Description |
+|---|---|--------|---------|
+| `name` | `str` | `None` | The model ID of the model to use. Supported base models for Text Embeddings: voyage-4, voyage-4-lite, voyage-4-large, voyage-3, voyage-3-lite, voyage-finance-2, voyage-multilingual-2, voyage-law-2, voyage-code-2 |
+| `input_type` | `str` | `None` | Type of the input text. Default to None. Other options: query, document. |
+| `truncation` | `bool` | `True` | Whether to truncate the input texts to fit within the context length. |
+
+
+Usage Example:
+    
+```python
+    import lancedb
+    from lancedb.pydantic import LanceModel, Vector
+    from lancedb.embeddings import EmbeddingFunctionRegistry
+
+    voyageai = EmbeddingFunctionRegistry
+        .get_instance()
+        .get("voyageai")
+        .create(name="voyage-3")
+
+    class TextModel(LanceModel):
+        text: str = voyageai.SourceField()
+        vector: Vector(voyageai.ndims()) =  voyageai.VectorField()
+
+    data = [ { "text": "hello world" },
+            { "text": "goodbye world" }]
+
+    db = lancedb.connect("~/.lancedb")
+    tbl = db.create_table("test", schema=TextModel, mode="overwrite")
+
+    tbl.add(data)
+```

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>0.23.1</version>
+    <version>0.26.1</version>
 </dependency>
 ```
 

docs/src/js/classes/Table.md

@@ -367,6 +367,27 @@ Use [Table.listIndices](Table.md#listindices) to find the names of the indices.
 
 ***
 
+### initialStorageOptions()
+
+```ts
+abstract initialStorageOptions(): Promise<undefined | null | Record<string, string>>
+```
+
+Get the initial storage options that were passed in when opening this table.
+
+For dynamically refreshed options (e.g., credential vending), use
+[Table.latestStorageOptions](Table.md#lateststorageoptions).
+
+Warning: This is an internal API and the return value is subject to change.
+
+#### Returns
+
+`Promise`&lt;`undefined` \| `null` \| `Record`&lt;`string`, `string`&gt;&gt;
+
+The storage options, or undefined if no storage options were configured.
+
+***
+
 ### isOpen()
 
 ```ts
@@ -381,6 +402,28 @@ Return true if the table has not been closed
 
 ***
 
+### latestStorageOptions()
+
+```ts
+abstract latestStorageOptions(): Promise<undefined | null | Record<string, string>>
+```
+
+Get the latest storage options, refreshing from provider if configured.
+
+This method is useful for credential vending scenarios where storage options
+may be refreshed dynamically. If no dynamic provider is configured, this
+returns the initial static options.
+
+Warning: This is an internal API and the return value is subject to change.
+
+#### Returns
+
+`Promise`&lt;`undefined` \| `null` \| `Record`&lt;`string`, `string`&gt;&gt;
+
+The storage options, or undefined if no storage options were configured.
+
+***
+
 ### listIndices()
 
 ```ts
@@ -705,8 +748,11 @@ Create a query that returns a subset of the rows in the table.
 
 #### Parameters
 
-* **rowIds**: `number`[]
+* **rowIds**: readonly (`number` \| `bigint`)[]
     The row ids of the rows to return.
+    Row ids returned by `withRowId()` are `bigint`, so `bigint[]` is supported.
+    For convenience / backwards compatibility, `number[]` is also accepted (for
+    small row ids that fit in a safe integer).
 
 #### Returns
 

java/lancedb-core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
       <groupId>com.lancedb</groupId>
       <artifactId>lancedb-parent</artifactId>
-      <version>0.23.1-final.0</version>
+      <version>0.26.1-final.0</version>
       <relativePath>../pom.xml</relativePath>
     </parent>
 

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.23.1-final.0</version>
+    <version>0.26.1-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>1.0.0-rc.2</lance-core.version>
+        <lance-core.version>2.0.0</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>
@@ -292,11 +292,12 @@
                     <plugin>
                         <groupId>org.sonatype.central</groupId>
                         <artifactId>central-publishing-maven-plugin</artifactId>
-                        <version>0.4.0</version>
+                        <version>0.8.0</version>
                         <extensions>true</extensions>
                         <configuration>
                             <publishingServerId>ossrh</publishingServerId>
                             <tokenAuth>true</tokenAuth>
+                            <autoPublish>true</autoPublish>
                         </configuration>
                     </plugin>
                     <plugin>

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.23.1"
+version = "0.26.1"
 license.workspace = true
 description.workspace = true
 repository.workspace = true

nodejs/__test__/table.test.ts

@@ -312,6 +312,66 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(res.getChild("id")?.toJSON()).toEqual([2, 3]);
     });
 
+    it("should support takeRowIds with bigint array", async () => {
+      await table.add([{ id: 1 }, { id: 2 }, { id: 3 }]);
+      // Get actual row IDs using withRowId()
+      const allRows = await table.query().withRowId().toArray();
+      const rowIds = allRows.map((row) => row._rowid) as bigint[];
+
+      // Verify row IDs are bigint
+      expect(typeof rowIds[0]).toBe("bigint");
+
+      // Use takeRowIds with bigint array (the main use case from issue #2722)
+      const res = await table.takeRowIds([rowIds[0], rowIds[2]]).toArray();
+      expect(res.map((r) => r.id)).toEqual([1, 3]);
+    });
+
+    it("should support takeRowIds with number array for backwards compatibility", async () => {
+      await table.add([{ id: 1 }, { id: 2 }, { id: 3 }]);
+      // Small row IDs can be passed as numbers
+      const res = await table.takeRowIds([0, 2]).toArray();
+      expect(res.map((r) => r.id)).toEqual([1, 3]);
+    });
+
+    it("should support takeRowIds with mixed bigint and number array", async () => {
+      await table.add([{ id: 1 }, { id: 2 }, { id: 3 }]);
+      // Mixed array of bigint and number
+      const res = await table.takeRowIds([0n, 1, 2n]).toArray();
+      expect(res.map((r) => r.id)).toEqual([1, 2, 3]);
+    });
+
+    it("should throw for non-integer number in takeRowIds", () => {
+      expect(() => table.takeRowIds([1.5])).toThrow(
+        "Row id must be an integer (or bigint)",
+      );
+      expect(() => table.takeRowIds([0, 1.1, 2])).toThrow(
+        "Row id must be an integer (or bigint)",
+      );
+    });
+
+    it("should throw for negative number in takeRowIds", () => {
+      expect(() => table.takeRowIds([-1])).toThrow("Row id cannot be negative");
+      expect(() => table.takeRowIds([0, -5, 2])).toThrow(
+        "Row id cannot be negative",
+      );
+    });
+
+    it("should throw for unsafe large number in takeRowIds", () => {
+      // Number.MAX_SAFE_INTEGER + 1 is not safe
+      const unsafeNumber = Number.MAX_SAFE_INTEGER + 1;
+      expect(() => table.takeRowIds([unsafeNumber])).toThrow(
+        "Row id is too large for number; use bigint instead",
+      );
+    });
+
+    it("should reject negative bigint in takeRowIds", async () => {
+      await table.add([{ id: 1 }]);
+      // Negative bigint should be rejected by the Rust layer
+      expect(() => {
+        table.takeRowIds([-1n]);
+      }).toThrow("Row id cannot be negative");
+    });
+
     it("should return the table as an instance of an arrow table", async () => {
       const arrowTbl = await table.toArrow();
       expect(arrowTbl).toBeInstanceOf(ArrowTable);
@@ -1520,9 +1580,9 @@ describe("when optimizing a dataset", () => {
 
   it("delete unverified", async () => {
     const version = await table.version();
-    const versionFile = `${tmpDir.name}/${table.name}.lance/_versions/${
-      version - 1
-    }.manifest`;
+    const versionFile = `${tmpDir.name}/${table.name}.lance/_versions/${String(
+      18446744073709551615n - (BigInt(version) - 1n),
+    ).padStart(20, "0")}.manifest`;
     fs.rmSync(versionFile);
 
     let stats = await table.optimize({ deleteUnverified: false });

nodejs/lancedb/table.ts

@@ -347,9 +347,13 @@ export abstract class Table {
   /**
    * Create a query that returns a subset of the rows in the table.
    * @param rowIds The row ids of the rows to return.
+   *
+   * Row ids returned by `withRowId()` are `bigint`, so `bigint[]` is supported.
+   * For convenience / backwards compatibility, `number[]` is also accepted (for
+   * small row ids that fit in a safe integer).
    * @returns A builder that can be used to parameterize the query.
    */
-  abstract takeRowIds(rowIds: number[]): TakeQuery;
+  abstract takeRowIds(rowIds: readonly (bigint | number)[]): TakeQuery;
 
   /**
    * Create a search query to find the nearest neighbors
@@ -538,6 +542,35 @@ export abstract class Table {
    *
    */
   abstract stats(): Promise<TableStatistics>;
+
+  /**
+   * Get the initial storage options that were passed in when opening this table.
+   *
+   * For dynamically refreshed options (e.g., credential vending), use
+   * {@link Table.latestStorageOptions}.
+   *
+   * Warning: This is an internal API and the return value is subject to change.
+   *
+   * @returns The storage options, or undefined if no storage options were configured.
+   */
+  abstract initialStorageOptions(): Promise<
+    Record<string, string> | null | undefined
+  >;
+
+  /**
+   * Get the latest storage options, refreshing from provider if configured.
+   *
+   * This method is useful for credential vending scenarios where storage options
+   * may be refreshed dynamically. If no dynamic provider is configured, this
+   * returns the initial static options.
+   *
+   * Warning: This is an internal API and the return value is subject to change.
+   *
+   * @returns The storage options, or undefined if no storage options were configured.
+   */
+  abstract latestStorageOptions(): Promise<
+    Record<string, string> | null | undefined
+  >;
 }
 
 export class LocalTable extends Table {
@@ -686,8 +719,24 @@ export class LocalTable extends Table {
     return new TakeQuery(this.inner.takeOffsets(offsets));
   }
 
-  takeRowIds(rowIds: number[]): TakeQuery {
-    return new TakeQuery(this.inner.takeRowIds(rowIds));
+  takeRowIds(rowIds: readonly (bigint | number)[]): TakeQuery {
+    const ids = rowIds.map((id) => {
+      if (typeof id === "bigint") {
+        return id;
+      }
+      if (!Number.isInteger(id)) {
+        throw new Error("Row id must be an integer (or bigint)");
+      }
+      if (id < 0) {
+        throw new Error("Row id cannot be negative");
+      }
+      if (!Number.isSafeInteger(id)) {
+        throw new Error("Row id is too large for number; use bigint instead");
+      }
+      return BigInt(id);
+    });
+
+    return new TakeQuery(this.inner.takeRowIds(ids));
   }
 
   query(): Query {
@@ -858,6 +907,18 @@ export class LocalTable extends Table {
     return await this.inner.stats();
   }
 
+  async initialStorageOptions(): Promise<
+    Record<string, string> | null | undefined
+  > {
+    return await this.inner.initialStorageOptions();
+  }
+
+  async latestStorageOptions(): Promise<
+    Record<string, string> | null | undefined
+  > {
+    return await this.inner.latestStorageOptions();
+  }
+
   mergeInsert(on: string | string[]): MergeInsertBuilder {
     on = Array.isArray(on) ? on : [on];
     return new MergeInsertBuilder(this.inner.mergeInsert(on), this.schema());

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",
@@ -8,5 +8,9 @@
 	"license": "Apache-2.0",
 	"engines": {
 		"node": ">= 18"
+	},
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
 	}
 }

nodejs/npm/darwin-x64/README.md

@@ -1,3 +0,0 @@
-# `@lancedb/lancedb-darwin-x64`
-
-This is the **x86_64-apple-darwin** binary for `@lancedb/lancedb`

nodejs/npm/darwin-x64/package.json

@@ -1,12 +0,0 @@
-{
-	"name": "@lancedb/lancedb-darwin-x64",
-	"version": "0.23.1",
-	"os": ["darwin"],
-	"cpu": ["x64"],
-	"main": "lancedb.darwin-x64.node",
-	"files": ["lancedb.darwin-x64.node"],
-	"license": "Apache-2.0",
-	"engines": {
-		"node": ">= 18"
-	}
-}

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",
@@ -9,5 +9,9 @@
 	"engines": {
 		"node": ">= 18"
 	},
-	"libc": ["glibc"]
+	"libc": ["glibc"],
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
+	}
 }

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",
@@ -9,5 +9,9 @@
 	"engines": {
 		"node": ">= 18"
 	},
-	"libc": ["musl"]
+	"libc": ["musl"],
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
+	}
 }

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",
@@ -9,5 +9,9 @@
 	"engines": {
 		"node": ">= 18"
 	},
-	"libc": ["glibc"]
+	"libc": ["glibc"],
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
+	}
 }

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",
@@ -9,5 +9,9 @@
 	"engines": {
 		"node": ">= 18"
 	},
-	"libc": ["musl"]
+	"libc": ["musl"],
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
+	}
 }

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

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.23.1",
+  "version": "0.26.1",
   "os": [
     "win32"
   ],
@@ -14,5 +14,9 @@
   "license": "Apache-2.0",
   "engines": {
     "node": ">= 18"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/lancedb/lancedb"
   }
 }

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.23.1",
+	"version": "0.26.1",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",
@@ -8,5 +8,9 @@
 	"license": "Apache-2.0",
 	"engines": {
 		"node": ">= 18"
+	},
+	"repository": {
+		"type": "git",
+		"url": "https://github.com/lancedb/lancedb"
 	}
 }

nodejs/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@lancedb/lancedb",
-  "version": "0.23.1",
+  "version": "0.26.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@lancedb/lancedb",
-      "version": "0.23.1",
+      "version": "0.26.1",
       "cpu": [
         "x64",
         "arm64"

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.23.1",
+  "version": "0.26.1",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",
@@ -25,7 +25,6 @@
     "triples": {
       "defaults": false,
       "additional": [
-        "x86_64-apple-darwin",
         "aarch64-apple-darwin",
         "x86_64-unknown-linux-gnu",
         "aarch64-unknown-linux-gnu",
@@ -37,6 +36,10 @@
     }
   },
   "license": "Apache-2.0",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/lancedb/lancedb"
+  },
   "devDependencies": {
     "@aws-sdk/client-dynamodb": "^3.33.0",
     "@aws-sdk/client-kms": "^3.33.0",

nodejs/src/table.rs

@@ -166,6 +166,19 @@ impl Table {
         Ok(stats.into())
     }
 
+    #[napi(catch_unwind)]
+    pub async fn initial_storage_options(&self) -> napi::Result<Option<HashMap<String, String>>> {
+        Ok(self.inner_ref()?.initial_storage_options().await)
+    }
+
+    #[napi(catch_unwind)]
+    pub async fn latest_storage_options(&self) -> napi::Result<Option<HashMap<String, String>>> {
+        self.inner_ref()?
+            .latest_storage_options()
+            .await
+            .default_error()
+    }
+
     #[napi(catch_unwind)]
     pub async fn update(
         &self,
@@ -208,18 +221,24 @@ impl Table {
     }
 
     #[napi(catch_unwind)]
-    pub fn take_row_ids(&self, row_ids: Vec<i64>) -> napi::Result<TakeQuery> {
+    pub fn take_row_ids(&self, row_ids: Vec<BigInt>) -> napi::Result<TakeQuery> {
         Ok(TakeQuery::new(
             self.inner_ref()?.take_row_ids(
                 row_ids
                     .into_iter()
-                    .map(|o| {
-                        u64::try_from(o).map_err(|e| {
-                            napi::Error::from_reason(format!(
-                                "Failed to convert row id to u64: {}",
-                                e
+                    .map(|id| {
+                        let (negative, value, lossless) = id.get_u64();
+                        if negative {
+                            Err(napi::Error::from_reason(
+                                "Row id cannot be negative".to_string(),
                             ))
-                        })
+                        } else if !lossless {
+                            Err(napi::Error::from_reason(
+                                "Row id is too large to fit in u64".to_string(),
+                            ))
+                        } else {
+                            Ok(value)
+                        }
                     })
                     .collect::<Result<Vec<_>>>()?,
             ),

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.27.0-beta.0"
+current_version = "0.29.2"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/CONTRIBUTING.md

@@ -16,7 +16,7 @@ The Python package is a wrapper around the Rust library, `lancedb`. We use
 
 To set up your development environment, you will need to install the following:
 
-1. Python 3.9 or later
+1. Python 3.10 or later
 2. Cargo (Rust's package manager). Use [rustup](https://rustup.rs/) to install.
 3. [protoc](https://grpc.io/docs/protoc-installation/) (Protocol Buffers compiler)
 

python/Cargo.toml

@@ -1,13 +1,13 @@
 [package]
 name = "lancedb-python"
-version = "0.27.0-beta.0"
+version = "0.29.2"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true
 repository.workspace = true
 keywords.workspace = true
 categories.workspace = true
-rust-version = "1.75.0"
+rust-version = "1.88.0"
 
 [lib]
 name = "_lancedb"

python/pyproject.toml

@@ -16,7 +16,7 @@ description = "lancedb"
 authors = [{ name = "LanceDB Devs", email = "dev@lancedb.com" }]
 license = { file = "LICENSE" }
 readme = "README.md"
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 keywords = [
     "data-format",
     "data-science",
@@ -33,10 +33,10 @@ classifiers = [
     "Programming Language :: Python",
     "Programming Language :: Python :: 3",
     "Programming Language :: Python :: 3 :: Only",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
     "Topic :: Scientific/Engineering",
 ]
 
@@ -137,4 +137,4 @@ include = [
     "python/lancedb/_lancedb.pyi",
 ]
 exclude = ["python/tests/"]
-pythonVersion = "3.12"
+pythonVersion = "3.13"

python/python/lancedb/_lancedb.pyi

@@ -180,6 +180,8 @@ class Table:
         delete_unverified: Optional[bool] = None,
     ) -> OptimizeStats: ...
     async def uri(self) -> str: ...
+    async def initial_storage_options(self) -> Optional[Dict[str, str]]: ...
+    async def latest_storage_options(self) -> Optional[Dict[str, str]]: ...
     @property
     def tags(self) -> Tags: ...
     def query(self) -> Query: ...

python/python/lancedb/background_loop.py

@@ -22,7 +22,12 @@ class BackgroundEventLoop:
         self.thread.start()
 
     def run(self, future):
-        return asyncio.run_coroutine_threadsafe(future, self.loop).result()
+        concurrent_future = asyncio.run_coroutine_threadsafe(future, self.loop)
+        try:
+            return concurrent_future.result()
+        except BaseException:
+            concurrent_future.cancel()
+            raise
 
 
 LOOP = BackgroundEventLoop()

python/python/lancedb/embeddings/colpali.py

@@ -275,7 +275,7 @@ class ColPaliEmbeddings(EmbeddingFunction):
         """
         Convert image inputs to PIL Images.
         """
-        PIL = attempt_import_or_raise("PIL", "pillow")
+        PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
         requests = attempt_import_or_raise("requests", "requests")
         images = self.sanitize_input(images)
         pil_images = []
@@ -285,12 +285,12 @@ class ColPaliEmbeddings(EmbeddingFunction):
                     if image.startswith(("http://", "https://")):
                         response = requests.get(image, timeout=10)
                         response.raise_for_status()
-                        pil_images.append(PIL.Image.open(io.BytesIO(response.content)))
+                        pil_images.append(PIL_Image.open(io.BytesIO(response.content)))
                     else:
-                        with PIL.Image.open(image) as im:
+                        with PIL_Image.open(image) as im:
                             pil_images.append(im.copy())
                 elif isinstance(image, bytes):
-                    pil_images.append(PIL.Image.open(io.BytesIO(image)))
+                    pil_images.append(PIL_Image.open(io.BytesIO(image)))
                 else:
                     # Assume it's a PIL Image; will raise if invalid
                     pil_images.append(image)

python/python/lancedb/embeddings/jinaai.py

@@ -77,8 +77,8 @@ class JinaEmbeddings(EmbeddingFunction):
             if isinstance(inputs, list):
                 inputs = inputs
             else:
-                PIL = attempt_import_or_raise("PIL", "pillow")
-                if isinstance(inputs, PIL.Image.Image):
+                PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
+                if isinstance(inputs, PIL_Image.Image):
                     inputs = [inputs]
         return inputs
 
@@ -89,13 +89,13 @@ class JinaEmbeddings(EmbeddingFunction):
         elif isinstance(image, (str, Path)):
             parsed = urlparse.urlparse(image)
             # TODO handle drive letter on windows.
-            PIL = attempt_import_or_raise("PIL", "pillow")
+            PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
             if parsed.scheme == "file":
-                pil_image = PIL.Image.open(parsed.path)
+                pil_image = PIL_Image.open(parsed.path)
             elif parsed.scheme == "":
-                pil_image = PIL.Image.open(image if os.name == "nt" else parsed.path)
+                pil_image = PIL_Image.open(image if os.name == "nt" else parsed.path)
             elif parsed.scheme.startswith("http"):
-                pil_image = PIL.Image.open(io.BytesIO(url_retrieve(image)))
+                pil_image = PIL_Image.open(io.BytesIO(url_retrieve(image)))
             else:
                 raise NotImplementedError("Only local and http(s) urls are supported")
             buffered = io.BytesIO()
@@ -103,9 +103,9 @@ class JinaEmbeddings(EmbeddingFunction):
             image_bytes = buffered.getvalue()
             image_dict = {"image": base64.b64encode(image_bytes).decode("utf-8")}
         else:
-            PIL = attempt_import_or_raise("PIL", "pillow")
+            PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
 
-            if isinstance(image, PIL.Image.Image):
+            if isinstance(image, PIL_Image.Image):
                 buffered = io.BytesIO()
                 image.save(buffered, format="PNG")
                 image_bytes = buffered.getvalue()
@@ -136,9 +136,9 @@ class JinaEmbeddings(EmbeddingFunction):
         elif isinstance(query, (Path, bytes)):
             return [self.generate_image_embedding(query)]
         else:
-            PIL = attempt_import_or_raise("PIL", "pillow")
+            PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
 
-            if isinstance(query, PIL.Image.Image):
+            if isinstance(query, PIL_Image.Image):
                 return [self.generate_image_embedding(query)]
             else:
                 raise TypeError(

python/python/lancedb/embeddings/open_clip.py

@@ -71,8 +71,8 @@ class OpenClipEmbeddings(EmbeddingFunction):
         if isinstance(query, str):
             return [self.generate_text_embeddings(query)]
         else:
-            PIL = attempt_import_or_raise("PIL", "pillow")
-            if isinstance(query, PIL.Image.Image):
+            PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
+            if isinstance(query, PIL_Image.Image):
                 return [self.generate_image_embedding(query)]
             else:
                 raise TypeError("OpenClip supports str or PIL Image as query")
@@ -145,20 +145,20 @@ class OpenClipEmbeddings(EmbeddingFunction):
             return self._encode_and_normalize_image(image)
 
     def _to_pil(self, image: Union[str, bytes]):
-        PIL = attempt_import_or_raise("PIL", "pillow")
+        PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
         if isinstance(image, bytes):
-            return PIL.Image.open(io.BytesIO(image))
-        if isinstance(image, PIL.Image.Image):
+            return PIL_Image.open(io.BytesIO(image))
+        if isinstance(image, PIL_Image.Image):
             return image
         elif isinstance(image, str):
             parsed = urlparse.urlparse(image)
             # TODO handle drive letter on windows.
             if parsed.scheme == "file":
-                return PIL.Image.open(parsed.path)
+                return PIL_Image.open(parsed.path)
             elif parsed.scheme == "":
-                return PIL.Image.open(image if os.name == "nt" else parsed.path)
+                return PIL_Image.open(image if os.name == "nt" else parsed.path)
             elif parsed.scheme.startswith("http"):
-                return PIL.Image.open(io.BytesIO(url_retrieve(image)))
+                return PIL_Image.open(io.BytesIO(url_retrieve(image)))
             else:
                 raise NotImplementedError("Only local and http(s) urls are supported")
 

python/python/lancedb/embeddings/siglip.py

@@ -56,8 +56,8 @@ class SigLipEmbeddings(EmbeddingFunction):
         if isinstance(query, str):
             return [self.generate_text_embeddings(query)]
         else:
-            PIL = attempt_import_or_raise("PIL", "pillow")
-            if isinstance(query, PIL.Image.Image):
+            PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
+            if isinstance(query, PIL_Image.Image):
                 return [self.generate_image_embedding(query)]
             else:
                 raise TypeError("SigLIP supports str or PIL Image as query")
@@ -127,21 +127,21 @@ class SigLipEmbeddings(EmbeddingFunction):
             return image_features.cpu().detach().numpy().squeeze()
 
     def _to_pil(self, image: Union[str, bytes, "PIL.Image.Image"]):
-        PIL = attempt_import_or_raise("PIL", "pillow")
-        if isinstance(image, PIL.Image.Image):
+        PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
+        if isinstance(image, PIL_Image.Image):
             return image.convert("RGB") if image.mode != "RGB" else image
         elif isinstance(image, bytes):
-            return PIL.Image.open(io.BytesIO(image)).convert("RGB")
+            return PIL_Image.open(io.BytesIO(image)).convert("RGB")
         elif isinstance(image, str):
             parsed = urlparse.urlparse(image)
             if parsed.scheme == "file":
-                return PIL.Image.open(parsed.path).convert("RGB")
+                return PIL_Image.open(parsed.path).convert("RGB")
             elif parsed.scheme == "":
                 path = image if os.name == "nt" else parsed.path
-                return PIL.Image.open(path).convert("RGB")
+                return PIL_Image.open(path).convert("RGB")
             elif parsed.scheme.startswith("http"):
                 image_bytes = url_retrieve(image)
-                return PIL.Image.open(io.BytesIO(image_bytes)).convert("RGB")
+                return PIL_Image.open(io.BytesIO(image_bytes)).convert("RGB")
             else:
                 raise NotImplementedError("Only local and http(s) urls are supported")
         else:

python/python/lancedb/embeddings/voyageai.py

@@ -21,6 +21,9 @@ if TYPE_CHECKING:
 
 # Token limits for different VoyageAI models
 VOYAGE_TOTAL_TOKEN_LIMITS = {
+    "voyage-4": 320_000,
+    "voyage-4-lite": 1_000_000,
+    "voyage-4-large": 120_000,
     "voyage-context-3": 32_000,
     "voyage-3.5-lite": 1_000_000,
     "voyage-3.5": 320_000,
@@ -61,7 +64,7 @@ def is_video_path(path: Path) -> bool:
 
 
 def transform_input(input_data: Union[str, bytes, Path]):
-    PIL = attempt_import_or_raise("PIL", "pillow")
+    PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
     if isinstance(input_data, str):
         if is_valid_url(input_data):
             if is_video_url(input_data):
@@ -70,7 +73,7 @@ def transform_input(input_data: Union[str, bytes, Path]):
                 content = {"type": "image_url", "image_url": input_data}
         else:
             content = {"type": "text", "text": input_data}
-    elif isinstance(input_data, PIL.Image.Image):
+    elif isinstance(input_data, PIL_Image.Image):
         buffered = BytesIO()
         input_data.save(buffered, format="JPEG")
         img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
@@ -79,7 +82,7 @@ def transform_input(input_data: Union[str, bytes, Path]):
             "image_base64": "data:image/jpeg;base64," + img_str,
         }
     elif isinstance(input_data, bytes):
-        img = PIL.Image.open(BytesIO(input_data))
+        img = PIL_Image.open(BytesIO(input_data))
         buffered = BytesIO()
         img.save(buffered, format="JPEG")
         img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
@@ -98,7 +101,7 @@ def transform_input(input_data: Union[str, bytes, Path]):
                 "video_base64": video_str,
             }
         else:
-            img = PIL.Image.open(input_data)
+            img = PIL_Image.open(input_data)
             buffered = BytesIO()
             img.save(buffered, format="JPEG")
             img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
@@ -116,8 +119,8 @@ def sanitize_multimodal_input(inputs: Union[TEXT, IMAGES]) -> List[Any]:
     """
     Sanitize the input to the embedding function.
     """
-    PIL = attempt_import_or_raise("PIL", "pillow")
-    if isinstance(inputs, (str, bytes, Path, PIL.Image.Image)):
+    PIL_Image = attempt_import_or_raise("PIL.Image", "pillow")
+    if isinstance(inputs, (str, bytes, Path, PIL_Image.Image)):
         inputs = [inputs]
     elif isinstance(inputs, list):
         pass  # Already a list, use as-is
@@ -130,7 +133,7 @@ def sanitize_multimodal_input(inputs: Union[TEXT, IMAGES]) -> List[Any]:
             f"Input type {type(inputs)} not allowed with multimodal model."
         )
 
-    if not all(isinstance(x, (str, bytes, Path, PIL.Image.Image)) for x in inputs):
+    if not all(isinstance(x, (str, bytes, Path, PIL_Image.Image)) for x in inputs):
         raise ValueError("Each input should be either str, bytes, Path or Image.")
 
     return [transform_input(i) for i in inputs]
@@ -167,6 +170,9 @@ class VoyageAIEmbeddingFunction(EmbeddingFunction):
     name: str
         The name of the model to use. List of acceptable models:
 
+            * voyage-4 (1024 dims, general-purpose and multilingual retrieval)
+            * voyage-4-lite (1024 dims, optimized for latency and cost)
+            * voyage-4-large (1024 dims, best retrieval quality)
             * voyage-context-3
             * voyage-3.5
             * voyage-3.5-lite
@@ -215,6 +221,9 @@ class VoyageAIEmbeddingFunction(EmbeddingFunction):
     _FLEXIBLE_DIM_MODELS: ClassVar[list] = ["voyage-multimodal-3.5"]
     _VALID_DIMENSIONS: ClassVar[list] = [256, 512, 1024, 2048]
     text_embedding_models: list = [
+        "voyage-4",
+        "voyage-4-lite",
+        "voyage-4-large",
         "voyage-3.5",
         "voyage-3.5-lite",
         "voyage-3",
@@ -252,6 +261,9 @@ class VoyageAIEmbeddingFunction(EmbeddingFunction):
         elif self.name == "voyage-code-2":
             return 1536
         elif self.name in [
+            "voyage-4",
+            "voyage-4-lite",
+            "voyage-4-large",
             "voyage-context-3",
             "voyage-3.5",
             "voyage-3.5-lite",

python/python/lancedb/pydantic.py

@@ -275,7 +275,7 @@ def _py_type_to_arrow_type(py_type: Type[Any], field: FieldInfo) -> pa.DataType:
         return pa.timestamp("us", tz=tz)
     elif getattr(py_type, "__origin__", None) in (list, tuple):
         child = py_type.__args__[0]
-        return pa.list_(_py_type_to_arrow_type(child, field))
+        return _pydantic_list_child_to_arrow(child, field)
     raise TypeError(
         f"Converting Pydantic type to Arrow Type: unsupported type {py_type}."
     )
@@ -298,12 +298,18 @@ else:
 
 
 def _pydantic_type_to_arrow_type(tp: Any, field: FieldInfo) -> pa.DataType:
+    def _safe_issubclass(candidate: Any, base: type) -> bool:
+        try:
+            return issubclass(candidate, base)
+        except TypeError:
+            return False
+
     if inspect.isclass(tp):
-        if issubclass(tp, pydantic.BaseModel):
+        if _safe_issubclass(tp, pydantic.BaseModel):
             # Struct
             fields = _pydantic_model_to_fields(tp)
             return pa.struct(fields)
-        if issubclass(tp, FixedSizeListMixin):
+        if _safe_issubclass(tp, FixedSizeListMixin):
             if getattr(tp, "is_multi_vector", lambda: False)():
                 return pa.list_(pa.list_(tp.value_arrow_type(), tp.dim()))
             # For regular Vector
@@ -311,45 +317,67 @@ def _pydantic_type_to_arrow_type(tp: Any, field: FieldInfo) -> pa.DataType:
     return _py_type_to_arrow_type(tp, field)
 
 
+def _pydantic_list_child_to_arrow(child: Any, field: FieldInfo) -> pa.DataType:
+    unwrapped = _unwrap_optional_annotation(child)
+    if unwrapped is not None:
+        return pa.list_(
+            pa.field("item", _pydantic_type_to_arrow_type(unwrapped, field), True)
+        )
+    return pa.list_(_pydantic_type_to_arrow_type(child, field))
+
+
+def _unwrap_optional_annotation(annotation: Any) -> Any | None:
+    if isinstance(annotation, (_GenericAlias, GenericAlias)):
+        origin = annotation.__origin__
+        args = annotation.__args__
+        if origin == Union:
+            non_none = [arg for arg in args if arg is not type(None)]
+            if len(non_none) == 1 and len(non_none) != len(args):
+                return non_none[0]
+    elif sys.version_info >= (3, 10) and isinstance(annotation, types.UnionType):
+        args = annotation.__args__
+        non_none = [arg for arg in args if arg is not type(None)]
+        if len(non_none) == 1 and len(non_none) != len(args):
+            return non_none[0]
+    return None
+
+
 def _pydantic_to_arrow_type(field: FieldInfo) -> pa.DataType:
     """Convert a Pydantic FieldInfo to Arrow DataType"""
+    unwrapped = _unwrap_optional_annotation(field.annotation)
+    if unwrapped is not None:
+        return _pydantic_type_to_arrow_type(unwrapped, field)
     if isinstance(field.annotation, (_GenericAlias, GenericAlias)):
         origin = field.annotation.__origin__
         args = field.annotation.__args__
 
         if origin is list:
             child = args[0]
-            return pa.list_(_py_type_to_arrow_type(child, field))
-        elif origin == Union:
-            if len(args) == 2 and args[1] is type(None):
-                return _pydantic_type_to_arrow_type(args[0], field)
-    elif sys.version_info >= (3, 10) and isinstance(field.annotation, types.UnionType):
-        args = field.annotation.__args__
-        if len(args) == 2:
-            for typ in args:
-                if typ is type(None):
-                    continue
-                return _py_type_to_arrow_type(typ, field)
+            return _pydantic_list_child_to_arrow(child, field)
     return _pydantic_type_to_arrow_type(field.annotation, field)
 
 
 def is_nullable(field: FieldInfo) -> bool:
     """Check if a Pydantic FieldInfo is nullable."""
+    if _unwrap_optional_annotation(field.annotation) is not None:
+        return True
     if isinstance(field.annotation, (_GenericAlias, GenericAlias)):
         origin = field.annotation.__origin__
         args = field.annotation.__args__
         if origin == Union:
-            if len(args) == 2 and args[1] is type(None):
+            if any(typ is type(None) for typ in args):
                 return True
     elif sys.version_info >= (3, 10) and isinstance(field.annotation, types.UnionType):
         args = field.annotation.__args__
         for typ in args:
             if typ is type(None):
                 return True
-    elif inspect.isclass(field.annotation) and issubclass(
-        field.annotation, FixedSizeListMixin
-    ):
-        return field.annotation.nullable()
+    elif inspect.isclass(field.annotation):
+        try:
+            if issubclass(field.annotation, FixedSizeListMixin):
+                return field.annotation.nullable()
+        except TypeError:
+            return False
     return False
 
 

python/python/lancedb/query.py

@@ -1433,6 +1433,19 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         self._bypass_vector_index = True
         return self
 
+    def fast_search(self) -> LanceVectorQueryBuilder:
+        """
+        Skip a flat search of unindexed data. This will improve
+        search performance but search results will not include unindexed data.
+
+        Returns
+        -------
+        LanceVectorQueryBuilder
+            The LanceVectorQueryBuilder object.
+        """
+        self._fast_search = True
+        return self
+
 
 class LanceFtsQueryBuilder(LanceQueryBuilder):
     """A builder for full text search for LanceDB."""

python/python/lancedb/table.py

@@ -2222,6 +2222,37 @@ class LanceTable(Table):
     def uri(self) -> str:
         return LOOP.run(self._table.uri())
 
+    def initial_storage_options(self) -> Optional[Dict[str, str]]:
+        """Get the initial storage options that were passed in when opening this table.
+
+        For dynamically refreshed options (e.g., credential vending), use
+        :meth:`latest_storage_options`.
+
+        Warning: This is an internal API and the return value is subject to change.
+
+        Returns
+        -------
+        Optional[Dict[str, str]]
+            The storage options, or None if no storage options were configured.
+        """
+        return LOOP.run(self._table.initial_storage_options())
+
+    def latest_storage_options(self) -> Optional[Dict[str, str]]:
+        """Get the latest storage options, refreshing from provider if configured.
+
+        This method is useful for credential vending scenarios where storage options
+        may be refreshed dynamically. If no dynamic provider is configured, this
+        returns the initial static options.
+
+        Warning: This is an internal API and the return value is subject to change.
+
+        Returns
+        -------
+        Optional[Dict[str, str]]
+            The storage options, or None if no storage options were configured.
+        """
+        return LOOP.run(self._table.latest_storage_options())
+
     def create_scalar_index(
         self,
         column: str,
@@ -3624,6 +3655,37 @@ class AsyncTable:
         """
         return await self._inner.uri()
 
+    async def initial_storage_options(self) -> Optional[Dict[str, str]]:
+        """Get the initial storage options that were passed in when opening this table.
+
+        For dynamically refreshed options (e.g., credential vending), use
+        :meth:`latest_storage_options`.
+
+        Warning: This is an internal API and the return value is subject to change.
+
+        Returns
+        -------
+        Optional[Dict[str, str]]
+            The storage options, or None if no storage options were configured.
+        """
+        return await self._inner.initial_storage_options()
+
+    async def latest_storage_options(self) -> Optional[Dict[str, str]]:
+        """Get the latest storage options, refreshing from provider if configured.
+
+        This method is useful for credential vending scenarios where storage options
+        may be refreshed dynamically. If no dynamic provider is configured, this
+        returns the initial static options.
+
+        Warning: This is an internal API and the return value is subject to change.
+
+        Returns
+        -------
+        Optional[Dict[str, str]]
+            The storage options, or None if no storage options were configured.
+        """
+        return await self._inner.latest_storage_options()
+
     async def add(
         self,
         data: DATA,

python/python/tests/conftest.py

@@ -2,12 +2,27 @@
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 from datetime import timedelta
+
 from lancedb.db import AsyncConnection, DBConnection
 import lancedb
 import pytest
 import pytest_asyncio
 
 
+def pandas_string_type():
+    """Return the PyArrow string type that pandas uses for string columns.
+
+    pandas 3.0+ uses large_string for string columns, pandas 2.x uses string.
+    """
+    import pandas as pd
+    import pyarrow as pa
+
+    version = tuple(int(x) for x in pd.__version__.split(".")[:2])
+    if version >= (3, 0):
+        return pa.large_utf8()
+    return pa.utf8()
+
+
 # Use an in-memory database for most tests.
 @pytest.fixture
 def mem_db() -> DBConnection:

python/python/tests/test_db.py

@@ -268,6 +268,8 @@ async def test_create_table_from_iterator_async(mem_db_async: lancedb.AsyncConne
 
 
 def test_create_exist_ok(tmp_db: lancedb.DBConnection):
+    from conftest import pandas_string_type
+
     data = pd.DataFrame(
         {
             "vector": [[3.1, 4.1], [5.9, 26.5]],
@@ -286,10 +288,11 @@ def test_create_exist_ok(tmp_db: lancedb.DBConnection):
     assert tbl.schema == tbl2.schema
     assert len(tbl) == len(tbl2)
 
+    # pandas 3.0+ uses large_string, pandas 2.x uses string
     schema = pa.schema(
         [
             pa.field("vector", pa.list_(pa.float32(), list_size=2)),
-            pa.field("item", pa.utf8()),
+            pa.field("item", pandas_string_type()),
             pa.field("price", pa.float64()),
         ]
     )
@@ -299,7 +302,7 @@ def test_create_exist_ok(tmp_db: lancedb.DBConnection):
     bad_schema = pa.schema(
         [
             pa.field("vector", pa.list_(pa.float32(), list_size=2)),
-            pa.field("item", pa.utf8()),
+            pa.field("item", pandas_string_type()),
             pa.field("price", pa.float64()),
             pa.field("extra", pa.float32()),
         ]
@@ -365,6 +368,8 @@ async def test_create_mode_async(tmp_db_async: lancedb.AsyncConnection):
 
 @pytest.mark.asyncio
 async def test_create_exist_ok_async(tmp_db_async: lancedb.AsyncConnection):
+    from conftest import pandas_string_type
+
     data = pd.DataFrame(
         {
             "vector": [[3.1, 4.1], [5.9, 26.5]],
@@ -382,10 +387,11 @@ async def test_create_exist_ok_async(tmp_db_async: lancedb.AsyncConnection):
     assert tbl.name == tbl2.name
     assert await tbl.schema() == await tbl2.schema()
 
+    # pandas 3.0+ uses large_string, pandas 2.x uses string
     schema = pa.schema(
         [
             pa.field("vector", pa.list_(pa.float32(), list_size=2)),
-            pa.field("item", pa.utf8()),
+            pa.field("item", pandas_string_type()),
             pa.field("price", pa.float64()),
         ]
     )
@@ -595,6 +601,8 @@ def test_open_table_sync(tmp_db: lancedb.DBConnection):
 
 @pytest.mark.asyncio
 async def test_open_table(tmp_path):
+    from conftest import pandas_string_type
+
     db = await lancedb.connect_async(tmp_path)
     data = pd.DataFrame(
         {
@@ -614,10 +622,11 @@ async def test_open_table(tmp_path):
         )
         is not None
     )
+    # pandas 3.0+ uses large_string, pandas 2.x uses string
     assert await tbl.schema() == pa.schema(
         {
             "vector": pa.list_(pa.float32(), list_size=2),
-            "item": pa.utf8(),
+            "item": pandas_string_type(),
             "price": pa.float64(),
         }
     )

python/python/tests/test_embeddings_slow.py

@@ -517,19 +517,36 @@ def test_ollama_embedding(tmp_path):
 @pytest.mark.skipif(
     os.environ.get("VOYAGE_API_KEY") is None, reason="VOYAGE_API_KEY not set"
 )
-def test_voyageai_embedding_function():
-    voyageai = get_registry().get("voyageai").create(name="voyage-3", max_retries=0)
+@pytest.mark.parametrize(
+    "model_name,expected_dims",
+    [
+        ("voyage-3", 1024),
+        ("voyage-4", 1024),
+        ("voyage-4-lite", 1024),
+        ("voyage-4-large", 1024),
+    ],
+)
+def test_voyageai_embedding_function(model_name, expected_dims, tmp_path):
+    """Integration test for VoyageAI text embedding models with real API calls."""
+    voyageai = get_registry().get("voyageai").create(name=model_name, max_retries=0)
 
     class TextModel(LanceModel):
         text: str = voyageai.SourceField()
         vector: Vector(voyageai.ndims()) = voyageai.VectorField()
 
     df = pd.DataFrame({"text": ["hello world", "goodbye world"]})
-    db = lancedb.connect("~/lancedb")
+    db = lancedb.connect(tmp_path)
     tbl = db.create_table("test", schema=TextModel, mode="overwrite")
 
     tbl.add(df)
     assert len(tbl.to_pandas()["vector"][0]) == voyageai.ndims()
+    assert voyageai.ndims() == expected_dims, (
+        f"{model_name} should have {expected_dims} dimensions"
+    )
+
+    # Test search functionality
+    result = tbl.search("hello").limit(1).to_pandas()
+    assert result["text"][0] == "hello world"
 
 
 @pytest.mark.slow

python/python/tests/test_namespace_integration.py

@@ -26,6 +26,8 @@ import pytest
 from lance_namespace import (
     CreateEmptyTableRequest,
     CreateEmptyTableResponse,
+    DeclareTableRequest,
+    DeclareTableResponse,
     DescribeTableRequest,
     DescribeTableResponse,
     LanceNamespace,
@@ -160,6 +162,19 @@ class TrackingNamespace(LanceNamespace):
 
         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:

python/python/tests/test_permutation.py

@@ -438,11 +438,15 @@ def test_filter_with_splits(mem_db):
     row_count = permutation_tbl.count_rows()
     assert row_count == 67
 
-    data = permutation_tbl.search(None).to_arrow().to_pydict()
+    # Verify the permutation table only contains row_id and split_id
+    assert set(permutation_tbl.schema.names) == {"row_id", "split_id"}
+
+    row_ids = permutation_tbl.search(None).to_arrow().to_pydict()["row_id"]
+    data = tbl.take_row_ids(row_ids).to_arrow().to_pydict()
     categories = data["category"]
 
     # All categories should be A or B
-    assert all(cat in ["A", "B"] for cat in categories)
+    assert all(cat in ("A", "B") for cat in categories)
 
 
 def test_filter_with_shuffle(mem_db):

python/python/tests/test_pydantic.py

@@ -2,7 +2,6 @@
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 import json
-import sys
 from datetime import date, datetime
 from typing import List, Optional, Tuple
 
@@ -20,10 +19,6 @@ from pydantic import BaseModel
 from pydantic import Field
 
 
-@pytest.mark.skipif(
-    sys.version_info < (3, 9),
-    reason="using native type alias requires python3.9 or higher",
-)
 def test_pydantic_to_arrow():
     class StructModel(pydantic.BaseModel):
         a: str
@@ -83,10 +78,6 @@ def test_pydantic_to_arrow():
     assert schema == expect_schema
 
 
-@pytest.mark.skipif(
-    sys.version_info < (3, 10),
-    reason="using | type syntax requires python3.10 or higher",
-)
 def test_optional_types_py310():
     class TestModel(pydantic.BaseModel):
         a: str | None
@@ -105,10 +96,233 @@ def test_optional_types_py310():
     assert schema == expect_schema
 
 
-@pytest.mark.skipif(
-    sys.version_info > (3, 8),
-    reason="using native type alias requires python3.9 or higher",
-)
+def test_optional_structs():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        split: SplitInfo | None = None
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "split",
+                pa.struct(
+                    [
+                        pa.field("start_frame", pa.int64(), False),
+                        pa.field("end_frame", pa.int64(), False),
+                    ]
+                ),
+                True,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_optional_struct_list_py310():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: list[SplitInfo] | None = None
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.struct(
+                        [
+                            pa.field("start_frame", pa.int64(), False),
+                            pa.field("end_frame", pa.int64(), False),
+                        ]
+                    )
+                ),
+                True,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_nested_struct_list():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: list[SplitInfo]
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.struct(
+                        [
+                            pa.field("start_frame", pa.int64(), False),
+                            pa.field("end_frame", pa.int64(), False),
+                        ]
+                    )
+                ),
+                False,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_nested_struct_list_optional():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: Optional[list[SplitInfo]] = None
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.struct(
+                        [
+                            pa.field("start_frame", pa.int64(), False),
+                            pa.field("end_frame", pa.int64(), False),
+                        ]
+                    )
+                ),
+                True,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_nested_struct_list_optional_items():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: list[Optional[SplitInfo]]
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.field(
+                        "item",
+                        pa.struct(
+                            [
+                                pa.field("start_frame", pa.int64(), False),
+                                pa.field("end_frame", pa.int64(), False),
+                            ]
+                        ),
+                        True,
+                    )
+                ),
+                False,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_nested_struct_list_optional_container_and_items():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: Optional[list[Optional[SplitInfo]]] = None
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.field(
+                        "item",
+                        pa.struct(
+                            [
+                                pa.field("start_frame", pa.int64(), False),
+                                pa.field("end_frame", pa.int64(), False),
+                            ]
+                        ),
+                        True,
+                    )
+                ),
+                True,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_nested_struct_list_optional_items_pep604():
+    class SplitInfo(pydantic.BaseModel):
+        start_frame: int
+        end_frame: int
+
+    class TestModel(pydantic.BaseModel):
+        id: str
+        splits: list[SplitInfo | None]
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.utf8(), False),
+            pa.field(
+                "splits",
+                pa.list_(
+                    pa.field(
+                        "item",
+                        pa.struct(
+                            [
+                                pa.field("start_frame", pa.int64(), False),
+                                pa.field("end_frame", pa.int64(), False),
+                            ]
+                        ),
+                        True,
+                    )
+                ),
+                False,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
 def test_pydantic_to_arrow_py38():
     class StructModel(pydantic.BaseModel):
         a: str

python/python/tests/test_query.py

@@ -1499,3 +1499,30 @@ def test_search_empty_table(mem_db):
     # Search on empty table should return empty results, not crash
     results = table.search([1.0, 2.0]).limit(5).to_list()
     assert results == []
+
+
+def test_fast_search(tmp_path):
+    db = lancedb.connect(tmp_path)
+
+    # Generate data matching the async test style
+    vectors = pa.FixedShapeTensorArray.from_numpy_ndarray(
+        np.random.rand(256, 32)
+    ).storage
+
+    table = db.create_table("test", pa.table({"vector": vectors}))
+
+    # FIX: Pass arguments directly instead of using 'config=IvfPq(...)'
+    table.create_index(vector_column_name="vector", num_partitions=1, num_sub_vectors=1)
+
+    # Add data to ensure table has enough segments/rows
+    table.add(pa.table({"vector": vectors}))
+
+    q = [1.0] * 32
+
+    # 1. Normal Search -> Should include "LanceScan" (Brute Force / Scan)
+    plan = table.search(q).explain_plan(True)
+    assert "LanceScan" in plan
+
+    # 2. Fast Search -> Should NOT include "LanceScan" (Uses Index)
+    plan = table.search(q).fast_search().explain_plan(True)
+    assert "LanceScan" not in plan

python/python/tests/test_remote_db.py

@@ -8,7 +8,7 @@ import http.server
 import json
 import threading
 import time
-from unittest.mock import MagicMock
+from unittest.mock import MagicMock, patch
 import uuid
 from packaging.version import Version
 
@@ -601,7 +601,6 @@ def test_head():
 def test_query_sync_minimal():
     def handler(body):
         assert body == {
-            "distance_type": "l2",
             "k": 10,
             "prefilter": True,
             "refine_factor": None,
@@ -685,7 +684,6 @@ def test_query_sync_maximal():
 def test_query_sync_nprobes():
     def handler(body):
         assert body == {
-            "distance_type": "l2",
             "k": 10,
             "prefilter": True,
             "fast_search": True,
@@ -715,7 +713,6 @@ def test_query_sync_nprobes():
 def test_query_sync_no_max_nprobes():
     def handler(body):
         assert body == {
-            "distance_type": "l2",
             "k": 10,
             "prefilter": True,
             "fast_search": True,
@@ -838,7 +835,6 @@ def test_query_sync_hybrid():
         else:
             # Vector query
             assert body == {
-                "distance_type": "l2",
                 "k": 42,
                 "prefilter": True,
                 "refine_factor": None,
@@ -1203,3 +1199,22 @@ async def test_header_provider_overrides_static_headers():
         extra_headers={"X-API-Key": "static-key", "X-Extra": "extra-value"},
     ) as db:
         await db.table_names()
+
+
+@pytest.mark.parametrize("exception", [KeyboardInterrupt, SystemExit, GeneratorExit])
+def test_background_loop_cancellation(exception):
+    """Test that BackgroundEventLoop.run() cancels the future on interrupt."""
+    from lancedb.background_loop import BackgroundEventLoop
+
+    mock_future = MagicMock()
+    mock_future.result.side_effect = exception()
+
+    with (
+        patch.object(BackgroundEventLoop, "__init__", return_value=None),
+        patch("asyncio.run_coroutine_threadsafe", return_value=mock_future),
+    ):
+        loop = BackgroundEventLoop()
+        loop.loop = MagicMock()
+        with pytest.raises(exception):
+            loop.run(None)
+        mock_future.cancel.assert_called_once()

python/python/tests/test_table.py

@@ -1880,8 +1880,13 @@ async def test_optimize_delete_unverified(tmp_db_async: AsyncConnection, tmp_pat
         ],
     )
     version = await table.version()
-    path = tmp_path / "test.lance" / "_versions" / f"{version - 1}.manifest"
+    assert version == 2
+
+    # By removing a manifest file, we make the data files we just inserted unverified
+    version_name = 18446744073709551615 - (version - 1)
+    path = tmp_path / "test.lance" / "_versions" / f"{version_name:020}.manifest"
     os.remove(path)
+
     stats = await table.optimize(delete_unverified=False)
     assert stats.prune.old_versions_removed == 0
     stats = await table.optimize(

python/python/tests/test_util.py

@@ -528,12 +528,19 @@ def test_sanitize_data(
         else:
             expected_schema = schema
     else:
+        from conftest import pandas_string_type
+
+        # polars uses large_string, pandas 3.0+ uses large_string, others use string
+        if isinstance(data, pl.DataFrame):
+            text_type = pa.large_utf8()
+        elif isinstance(data, pd.DataFrame):
+            text_type = pandas_string_type()
+        else:
+            text_type = pa.string()
         expected_schema = pa.schema(
             {
                 "id": pa.int64(),
-                "text": pa.large_utf8()
-                if isinstance(data, pl.DataFrame)
-                else pa.string(),
+                "text": text_type,
                 "vector": pa.list_(pa.float32(), 10),
             }
         )

python/python/tests/test_voyageai_embeddings.py

@@ -0,0 +1,108 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+"""Unit tests for VoyageAI embedding function.
+
+These tests verify model registration and configuration without requiring API calls.
+"""
+
+import pytest
+from unittest.mock import MagicMock, patch
+
+from lancedb.embeddings import get_registry
+
+
+@pytest.fixture(autouse=True)
+def reset_voyageai_client():
+    """Reset VoyageAI client before and after each test to avoid state pollution."""
+    from lancedb.embeddings.voyageai import VoyageAIEmbeddingFunction
+
+    VoyageAIEmbeddingFunction.client = None
+    yield
+    VoyageAIEmbeddingFunction.client = None
+
+
+class TestVoyageAIModelRegistration:
+    """Tests for VoyageAI model registration and configuration."""
+
+    @pytest.fixture
+    def mock_voyageai_client(self):
+        """Mock VoyageAI client to avoid API calls."""
+        with patch.dict("os.environ", {"VOYAGE_API_KEY": "test-key"}):
+            with patch("lancedb.embeddings.voyageai.attempt_import_or_raise") as mock:
+                mock_client = MagicMock()
+                mock_voyageai = MagicMock()
+                mock_voyageai.Client.return_value = mock_client
+                mock.return_value = mock_voyageai
+                yield mock_client
+
+    def test_voyageai_registered(self):
+        """Test that VoyageAI is registered in the embedding function registry."""
+        registry = get_registry()
+        assert registry.get("voyageai") is not None
+
+    @pytest.mark.parametrize(
+        "model_name,expected_dims",
+        [
+            # Voyage-4 series (all 1024 dims)
+            ("voyage-4", 1024),
+            ("voyage-4-lite", 1024),
+            ("voyage-4-large", 1024),
+            # Voyage-3 series
+            ("voyage-3", 1024),
+            ("voyage-3-lite", 512),
+            # Domain-specific models
+            ("voyage-finance-2", 1024),
+            ("voyage-multilingual-2", 1024),
+            ("voyage-law-2", 1024),
+            ("voyage-code-2", 1536),
+            # Multimodal
+            ("voyage-multimodal-3", 1024),
+        ],
+    )
+    def test_model_dimensions(self, model_name, expected_dims, mock_voyageai_client):
+        """Test that each model returns the correct dimensions."""
+        registry = get_registry()
+        func = registry.get("voyageai").create(name=model_name)
+        assert func.ndims() == expected_dims, (
+            f"Model {model_name} should have {expected_dims} dimensions"
+        )
+
+    def test_unsupported_model_raises_error(self, mock_voyageai_client):
+        """Test that unsupported models raise ValueError."""
+        registry = get_registry()
+        func = registry.get("voyageai").create(name="unsupported-model")
+        with pytest.raises(ValueError, match="not supported"):
+            func.ndims()
+
+    @pytest.mark.parametrize(
+        "model_name",
+        [
+            "voyage-4",
+            "voyage-4-lite",
+            "voyage-4-large",
+        ],
+    )
+    def test_voyage4_models_are_text_models(self, model_name, mock_voyageai_client):
+        """Test that voyage-4 models are classified as text models (not multimodal)."""
+        registry = get_registry()
+        func = registry.get("voyageai").create(name=model_name)
+        assert not func._is_multimodal_model(model_name), (
+            f"{model_name} should be a text model, not multimodal"
+        )
+
+    def test_voyage4_models_in_text_embedding_list(self, mock_voyageai_client):
+        """Test that voyage-4 models are in the text_embedding_models list."""
+        registry = get_registry()
+        func = registry.get("voyageai").create(name="voyage-4")
+        assert "voyage-4" in func.text_embedding_models
+        assert "voyage-4-lite" in func.text_embedding_models
+        assert "voyage-4-large" in func.text_embedding_models
+
+    def test_voyage4_models_not_in_multimodal_list(self, mock_voyageai_client):
+        """Test that voyage-4 models are NOT in the multimodal_embedding_models list."""
+        registry = get_registry()
+        func = registry.get("voyageai").create(name="voyage-4")
+        assert "voyage-4" not in func.multimodal_embedding_models
+        assert "voyage-4-lite" not in func.multimodal_embedding_models
+        assert "voyage-4-large" not in func.multimodal_embedding_models

python/src/arrow.rs

@@ -55,12 +55,12 @@ impl RecordBatchStream {
                 .next()
                 .await
                 .ok_or_else(|| PyStopAsyncIteration::new_err(""))?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = inner_next.infer_error()?.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| {
+                inner_next
+                    .infer_error()?
+                    .to_pyarrow(py)
+                    .map(|obj| obj.unbind())
+            })
         })
     }
 }

python/src/connection.rs

@@ -307,8 +307,7 @@ impl Connection {
                 ..Default::default()
             };
             let response = inner.list_namespaces(request).await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| -> PyResult<Py<PyDict>> {
+            Python::attach(|py| -> PyResult<Py<PyDict>> {
                 let dict = PyDict::new(py);
                 dict.set_item("namespaces", response.namespaces)?;
                 dict.set_item("page_token", response.page_token)?;
@@ -328,7 +327,8 @@ impl Connection {
         let py = self_.py();
         future_into_py(py, async move {
             use lance_namespace::models::CreateNamespaceRequest;
-            let mode_enum = mode.and_then(|m| match m.to_lowercase().as_str() {
+            // Mode is now a string field
+            let mode_str = mode.and_then(|m| match m.to_lowercase().as_str() {
                 "create" => Some("Create".to_string()),
                 "exist_ok" => Some("ExistOk".to_string()),
                 "overwrite" => Some("Overwrite".to_string()),
@@ -340,13 +340,12 @@ impl Connection {
                 } else {
                     Some(namespace)
                 },
-                mode: mode_enum,
+                mode: mode_str,
                 properties,
                 ..Default::default()
             };
             let response = inner.create_namespace(request).await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| -> PyResult<Py<PyDict>> {
+            Python::attach(|py| -> PyResult<Py<PyDict>> {
                 let dict = PyDict::new(py);
                 dict.set_item("properties", response.properties)?;
                 Ok(dict.unbind())
@@ -365,12 +364,13 @@ impl Connection {
         let py = self_.py();
         future_into_py(py, async move {
             use lance_namespace::models::DropNamespaceRequest;
-            let mode_enum = mode.and_then(|m| match m.to_uppercase().as_str() {
+            // Mode and Behavior are now string fields
+            let mode_str = mode.and_then(|m| match m.to_uppercase().as_str() {
                 "SKIP" => Some("Skip".to_string()),
                 "FAIL" => Some("Fail".to_string()),
                 _ => None,
             });
-            let behavior_enum = behavior.and_then(|b| match b.to_uppercase().as_str() {
+            let behavior_str = behavior.and_then(|b| match b.to_uppercase().as_str() {
                 "RESTRICT" => Some("Restrict".to_string()),
                 "CASCADE" => Some("Cascade".to_string()),
                 _ => None,
@@ -381,13 +381,12 @@ impl Connection {
                 } else {
                     Some(namespace)
                 },
-                mode: mode_enum,
-                behavior: behavior_enum,
+                mode: mode_str,
+                behavior: behavior_str,
                 ..Default::default()
             };
             let response = inner.drop_namespace(request).await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| -> PyResult<Py<PyDict>> {
+            Python::attach(|py| -> PyResult<Py<PyDict>> {
                 let dict = PyDict::new(py);
                 dict.set_item("properties", response.properties)?;
                 dict.set_item("transaction_id", response.transaction_id)?;
@@ -414,8 +413,7 @@ impl Connection {
                 ..Default::default()
             };
             let response = inner.describe_namespace(request).await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| -> PyResult<Py<PyDict>> {
+            Python::attach(|py| -> PyResult<Py<PyDict>> {
                 let dict = PyDict::new(py);
                 dict.set_item("properties", response.properties)?;
                 Ok(dict.unbind())
@@ -445,8 +443,7 @@ impl Connection {
                 ..Default::default()
             };
             let response = inner.list_tables(request).await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| -> PyResult<Py<PyDict>> {
+            Python::attach(|py| -> PyResult<Py<PyDict>> {
                 let dict = PyDict::new(py);
                 dict.set_item("tables", response.tables)?;
                 dict.set_item("page_token", response.page_token)?;

python/src/error.rs

@@ -40,34 +40,31 @@ impl<T> PythonErrorExt<T> for std::result::Result<T, LanceError> {
                     request_id,
                     source,
                     status_code,
-                } => {
-                    #[allow(deprecated)]
-                    Python::with_gil(|py| {
-                        let message = err.to_string();
-                        let http_err_cls = py
-                            .import(intern!(py, "lancedb.remote.errors"))?
-                            .getattr(intern!(py, "HttpError"))?;
-                        let err = http_err_cls.call1((
-                            message,
+                } => Python::attach(|py| {
+                    let message = err.to_string();
+                    let http_err_cls = py
+                        .import(intern!(py, "lancedb.remote.errors"))?
+                        .getattr(intern!(py, "HttpError"))?;
+                    let err = http_err_cls.call1((
+                        message,
+                        request_id,
+                        status_code.map(|s| s.as_u16()),
+                    ))?;
+
+                    if let Some(cause) = source.source() {
+                        // The HTTP error already includes the first cause. But
+                        // we can add the rest of the chain if there is any more.
+                        let cause_err = http_from_rust_error(
+                            py,
+                            cause,
                             request_id,
                             status_code.map(|s| s.as_u16()),
-                        ))?;
+                        )?;
+                        err.setattr(intern!(py, "__cause__"), cause_err)?;
+                    }
 
-                        if let Some(cause) = source.source() {
-                            // The HTTP error already includes the first cause. But
-                            // we can add the rest of the chain if there is any more.
-                            let cause_err = http_from_rust_error(
-                                py,
-                                cause,
-                                request_id,
-                                status_code.map(|s| s.as_u16()),
-                            )?;
-                            err.setattr(intern!(py, "__cause__"), cause_err)?;
-                        }
-
-                        Err(PyErr::from_value(err))
-                    })
-                }
+                    Err(PyErr::from_value(err))
+                }),
                 LanceError::Retry {
                     request_id,
                     request_failures,
@@ -78,37 +75,33 @@ impl<T> PythonErrorExt<T> for std::result::Result<T, LanceError> {
                     max_read_failures,
                     source,
                     status_code,
-                } =>
-                {
-                    #[allow(deprecated)]
-                    Python::with_gil(|py| {
-                        let cause_err = http_from_rust_error(
-                            py,
-                            source.as_ref(),
-                            request_id,
-                            status_code.map(|s| s.as_u16()),
-                        )?;
+                } => Python::attach(|py| {
+                    let cause_err = http_from_rust_error(
+                        py,
+                        source.as_ref(),
+                        request_id,
+                        status_code.map(|s| s.as_u16()),
+                    )?;
 
-                        let message = err.to_string();
-                        let retry_error_cls = py
-                            .import(intern!(py, "lancedb.remote.errors"))?
-                            .getattr("RetryError")?;
-                        let err = retry_error_cls.call1((
-                            message,
-                            request_id,
-                            *request_failures,
-                            *connect_failures,
-                            *read_failures,
-                            *max_request_failures,
-                            *max_connect_failures,
-                            *max_read_failures,
-                            status_code.map(|s| s.as_u16()),
-                        ))?;
+                    let message = err.to_string();
+                    let retry_error_cls = py
+                        .import(intern!(py, "lancedb.remote.errors"))?
+                        .getattr("RetryError")?;
+                    let err = retry_error_cls.call1((
+                        message,
+                        request_id,
+                        *request_failures,
+                        *connect_failures,
+                        *read_failures,
+                        *max_request_failures,
+                        *max_connect_failures,
+                        *max_read_failures,
+                        status_code.map(|s| s.as_u16()),
+                    ))?;
 
-                        err.setattr(intern!(py, "__cause__"), cause_err)?;
-                        Err(PyErr::from_value(err))
-                    })
-                }
+                    err.setattr(intern!(py, "__cause__"), cause_err)?;
+                    Err(PyErr::from_value(err))
+                }),
                 _ => self.runtime_error(),
             },
         }

python/src/header.rs

@@ -12,8 +12,7 @@ pub struct PyHeaderProvider {
 
 impl Clone for PyHeaderProvider {
     fn clone(&self) -> Self {
-        #[allow(deprecated)]
-        Python::with_gil(|py| Self {
+        Python::attach(|py| Self {
             provider: self.provider.clone_ref(py),
         })
     }
@@ -26,8 +25,7 @@ impl PyHeaderProvider {
 
     /// Get headers from the Python provider (internal implementation)
     fn get_headers_internal(&self) -> Result<HashMap<String, String>, String> {
-        #[allow(deprecated)]
-        Python::with_gil(|py| {
+        Python::attach(|py| {
             // Call the get_headers method
             let result = self.provider.call_method0(py, "get_headers");
 

python/src/permutation.rs

@@ -19,7 +19,7 @@ use pyo3::{
     exceptions::PyRuntimeError,
     pyclass, pymethods,
     types::{PyAnyMethods, PyDict, PyDictMethods, PyType},
-    Bound, Py, PyAny, PyRef, PyRefMut, PyResult, Python,
+    Bound, PyAny, PyRef, PyRefMut, PyResult, Python,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 
@@ -281,12 +281,7 @@ impl PyPermutationReader {
         let reader = slf.reader.clone();
         future_into_py(slf.py(), async move {
             let schema = reader.output_schema(selection).await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 

python/src/query.rs

@@ -29,7 +29,6 @@ use pyo3::types::PyList;
 use pyo3::types::{PyDict, PyString};
 use pyo3::Bound;
 use pyo3::IntoPyObject;
-use pyo3::Py;
 use pyo3::PyAny;
 use pyo3::PyRef;
 use pyo3::PyResult;
@@ -454,12 +453,7 @@ impl Query {
         let inner = self_.inner.clone();
         future_into_py(self_.py(), async move {
             let schema = inner.output_schema().await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 
@@ -538,12 +532,7 @@ impl TakeQuery {
         let inner = self_.inner.clone();
         future_into_py(self_.py(), async move {
             let schema = inner.output_schema().await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 
@@ -638,12 +627,7 @@ impl FTSQuery {
         let inner = self_.inner.clone();
         future_into_py(self_.py(), async move {
             let schema = inner.output_schema().await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 
@@ -822,12 +806,7 @@ impl VectorQuery {
         let inner = self_.inner.clone();
         future_into_py(self_.py(), async move {
             let schema = inner.output_schema().await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 

python/src/storage_options.rs

@@ -22,8 +22,7 @@ pub struct PyStorageOptionsProvider {
 
 impl Clone for PyStorageOptionsProvider {
     fn clone(&self) -> Self {
-        #[allow(deprecated)]
-        Python::with_gil(|py| Self {
+        Python::attach(|py| Self {
             inner: self.inner.clone_ref(py),
         })
     }
@@ -31,17 +30,14 @@ impl Clone for PyStorageOptionsProvider {
 
 impl PyStorageOptionsProvider {
     pub fn new(obj: Py<PyAny>) -> PyResult<Self> {
-        #[allow(deprecated)]
-        Python::with_gil(|py| {
+        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.clone_ref(py),
-            })
+            Ok(Self { inner: obj })
         })
     }
 }
@@ -64,8 +60,7 @@ impl StorageOptionsProvider for PyStorageOptionsProviderWrapper {
         let py_provider = self.py_provider.clone();
 
         tokio::task::spawn_blocking(move || {
-            #[allow(deprecated)]
-            Python::with_gil(|py| {
+            Python::attach(|py| {
                 // Call the Python fetch_storage_options method
                 let result = py_provider
                     .inner
@@ -124,8 +119,7 @@ impl StorageOptionsProvider for PyStorageOptionsProviderWrapper {
     }
 
     fn provider_id(&self) -> String {
-        #[allow(deprecated)]
-        Python::with_gil(|py| {
+        Python::attach(|py| {
             // Call provider_id() method on the Python object
             let obj = self.py_provider.inner.bind(py);
             obj.call_method0("provider_id")

python/src/table.rs

@@ -21,7 +21,7 @@ use pyo3::{
     exceptions::{PyKeyError, PyRuntimeError, PyValueError},
     pyclass, pymethods,
     types::{IntoPyDict, PyAnyMethods, PyDict, PyDictMethods},
-    Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
+    Bound, FromPyObject, PyAny, PyRef, PyResult, Python,
 };
 use pyo3_async_runtimes::tokio::future_into_py;
 
@@ -287,12 +287,7 @@ impl Table {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
             let schema = inner.schema().await.infer_error()?;
-            #[allow(deprecated)]
-            let py_obj: Py<PyAny> = Python::with_gil(|py| -> PyResult<Py<PyAny>> {
-                let bound = schema.to_pyarrow(py)?;
-                Ok(bound.unbind())
-            })?;
-            Ok(py_obj)
+            Python::attach(|py| schema.to_pyarrow(py).map(|obj| obj.unbind()))
         })
     }
 
@@ -442,8 +437,7 @@ impl Table {
         future_into_py(self_.py(), async move {
             let stats = inner.index_stats(&index_name).await.infer_error()?;
             if let Some(stats) = stats {
-                #[allow(deprecated)]
-                Python::with_gil(|py| {
+                Python::attach(|py| {
                     let dict = PyDict::new(py);
                     dict.set_item("num_indexed_rows", stats.num_indexed_rows)?;
                     dict.set_item("num_unindexed_rows", stats.num_unindexed_rows)?;
@@ -473,8 +467,7 @@ impl Table {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
             let stats = inner.stats().await.infer_error()?;
-            #[allow(deprecated)]
-            Python::with_gil(|py| {
+            Python::attach(|py| {
                 let dict = PyDict::new(py);
                 dict.set_item("total_bytes", stats.total_bytes)?;
                 dict.set_item("num_rows", stats.num_rows)?;
@@ -509,6 +502,20 @@ impl Table {
         future_into_py(self_.py(), async move { inner.uri().await.infer_error() })
     }
 
+    pub fn initial_storage_options(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        future_into_py(self_.py(), async move {
+            Ok(inner.initial_storage_options().await)
+        })
+    }
+
+    pub fn latest_storage_options(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        future_into_py(self_.py(), async move {
+            inner.latest_storage_options().await.infer_error()
+        })
+    }
+
     pub fn __repr__(&self) -> String {
         match &self.inner {
             None => format!("ClosedTable({})", self.name),
@@ -528,8 +535,7 @@ impl Table {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
             let versions = inner.list_versions().await.infer_error()?;
-            #[allow(deprecated)]
-            let versions_as_dict = Python::with_gil(|py| {
+            let versions_as_dict = Python::attach(|py| {
                 versions
                     .iter()
                     .map(|v| {
@@ -880,8 +886,7 @@ impl Tags {
             let tags = inner.tags().await.infer_error()?;
             let res = tags.list().await.infer_error()?;
 
-            #[allow(deprecated)]
-            Python::with_gil(|py| {
+            Python::attach(|py| {
                 let py_dict = PyDict::new(py);
                 for (key, contents) in res {
                     let value_dict = PyDict::new(py);

rust/lancedb/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.23.1"
+version = "0.26.1"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true
@@ -25,6 +25,7 @@ datafusion-catalog.workspace = true
 datafusion-common.workspace = true
 datafusion-execution.workspace = true
 datafusion-expr.workspace = true
+datafusion-physical-expr.workspace = true
 datafusion-physical-plan.workspace = true
 datafusion.workspace = true
 object_store = { workspace = true }

rust/lancedb/src/connection.rs

@@ -36,10 +36,42 @@ use crate::remote::{
 };
 use crate::table::{TableDefinition, WriteOptions};
 use crate::Table;
+use lance::io::ObjectStoreParams;
 pub use lance_encoding::version::LanceFileVersion;
 #[cfg(feature = "remote")]
 use lance_io::object_store::StorageOptions;
-use lance_io::object_store::StorageOptionsProvider;
+use lance_io::object_store::{StorageOptionsAccessor, StorageOptionsProvider};
+
+fn merge_storage_options(
+    store_params: &mut ObjectStoreParams,
+    pairs: impl IntoIterator<Item = (String, String)>,
+) {
+    let mut options = store_params.storage_options().cloned().unwrap_or_default();
+    for (key, value) in pairs {
+        options.insert(key, value);
+    }
+    let provider = store_params
+        .storage_options_accessor
+        .as_ref()
+        .and_then(|accessor| accessor.provider().cloned());
+    let accessor = if let Some(provider) = provider {
+        StorageOptionsAccessor::with_initial_and_provider(options, provider)
+    } else {
+        StorageOptionsAccessor::with_static_options(options)
+    };
+    store_params.storage_options_accessor = Some(Arc::new(accessor));
+}
+
+fn set_storage_options_provider(
+    store_params: &mut ObjectStoreParams,
+    provider: Arc<dyn StorageOptionsProvider>,
+) {
+    let accessor = match store_params.storage_options().cloned() {
+        Some(options) => StorageOptionsAccessor::with_initial_and_provider(options, provider),
+        None => StorageOptionsAccessor::with_provider(provider),
+    };
+    store_params.storage_options_accessor = Some(Arc::new(accessor));
+}
 
 /// A builder for configuring a [`Connection::table_names`] operation
 pub struct TableNamesBuilder {
@@ -219,8 +251,36 @@ impl CreateTableBuilder<false> {
     /// Execute the create table operation
     pub async fn execute(self) -> Result<Table> {
         let parent = self.parent.clone();
-        let table = parent.create_table(self.request).await?;
-        Ok(Table::new(table, parent))
+        let embedding_registry = self.embedding_registry.clone();
+        let request = self.into_request()?;
+        Ok(Table::new_with_embedding_registry(
+            parent.create_table(request).await?,
+            parent,
+            embedding_registry,
+        ))
+    }
+
+    fn into_request(self) -> Result<CreateTableRequest> {
+        if self.embeddings.is_empty() {
+            return Ok(self.request);
+        }
+
+        let CreateTableData::Empty(table_def) = self.request.data else {
+            unreachable!("CreateTableBuilder<false> should always have Empty data")
+        };
+
+        let schema = table_def.schema.clone();
+        let empty_batch = arrow_array::RecordBatch::new_empty(schema.clone());
+
+        let reader = Box::new(std::iter::once(Ok(empty_batch)).collect::<Vec<_>>());
+        let reader = arrow_array::RecordBatchIterator::new(reader.into_iter(), schema);
+        let with_embeddings = WithEmbeddings::new(reader, self.embeddings);
+        let table_definition = with_embeddings.table_definition()?;
+
+        Ok(CreateTableRequest {
+            data: CreateTableData::Empty(table_definition),
+            ..self.request
+        })
     }
 }
 
@@ -246,16 +306,14 @@ impl<const HAS_DATA: bool> CreateTableBuilder<HAS_DATA> {
     ///
     /// See available options at <https://lancedb.com/docs/storage/>
     pub fn storage_option(mut self, key: impl Into<String>, value: impl Into<String>) -> Self {
-        let store_options = self
+        let store_params = self
             .request
             .write_options
             .lance_write_params
             .get_or_insert(Default::default())
             .store_params
-            .get_or_insert(Default::default())
-            .storage_options
             .get_or_insert(Default::default());
-        store_options.insert(key.into(), value.into());
+        merge_storage_options(store_params, [(key.into(), value.into())]);
         self
     }
 
@@ -269,19 +327,17 @@ impl<const HAS_DATA: bool> CreateTableBuilder<HAS_DATA> {
         mut self,
         pairs: impl IntoIterator<Item = (impl Into<String>, impl Into<String>)>,
     ) -> Self {
-        let store_options = self
+        let store_params = self
             .request
             .write_options
             .lance_write_params
             .get_or_insert(Default::default())
             .store_params
-            .get_or_insert(Default::default())
-            .storage_options
             .get_or_insert(Default::default());
-
-        for (key, value) in pairs {
-            store_options.insert(key.into(), value.into());
-        }
+        let updates = pairs
+            .into_iter()
+            .map(|(key, value)| (key.into(), value.into()));
+        merge_storage_options(store_params, updates);
         self
     }
 
@@ -318,23 +374,21 @@ impl<const HAS_DATA: bool> CreateTableBuilder<HAS_DATA> {
     /// This has no effect in LanceDB Cloud.
     #[deprecated(since = "0.15.1", note = "Use `database_options` instead")]
     pub fn enable_v2_manifest_paths(mut self, use_v2_manifest_paths: bool) -> Self {
-        let storage_options = self
+        let store_params = self
             .request
             .write_options
             .lance_write_params
             .get_or_insert_with(Default::default)
             .store_params
-            .get_or_insert_with(Default::default)
-            .storage_options
             .get_or_insert_with(Default::default);
-
-        storage_options.insert(
-            OPT_NEW_TABLE_V2_MANIFEST_PATHS.to_string(),
-            if use_v2_manifest_paths {
-                "true".to_string()
-            } else {
-                "false".to_string()
-            },
+        let value = if use_v2_manifest_paths {
+            "true".to_string()
+        } else {
+            "false".to_string()
+        };
+        merge_storage_options(
+            store_params,
+            [(OPT_NEW_TABLE_V2_MANIFEST_PATHS.to_string(), value)],
         );
         self
     }
@@ -344,19 +398,19 @@ impl<const HAS_DATA: bool> CreateTableBuilder<HAS_DATA> {
     /// The default is `LanceFileVersion::Stable`.
     #[deprecated(since = "0.15.1", note = "Use `database_options` instead")]
     pub fn data_storage_version(mut self, data_storage_version: LanceFileVersion) -> Self {
-        let storage_options = self
+        let store_params = self
             .request
             .write_options
             .lance_write_params
             .get_or_insert_with(Default::default)
             .store_params
-            .get_or_insert_with(Default::default)
-            .storage_options
             .get_or_insert_with(Default::default);
-
-        storage_options.insert(
-            OPT_NEW_TABLE_STORAGE_VERSION.to_string(),
-            data_storage_version.to_string(),
+        merge_storage_options(
+            store_params,
+            [(
+                OPT_NEW_TABLE_STORAGE_VERSION.to_string(),
+                data_storage_version.to_string(),
+            )],
         );
         self
     }
@@ -381,13 +435,14 @@ impl<const HAS_DATA: bool> CreateTableBuilder<HAS_DATA> {
     /// This allows tables to automatically refresh cloud storage credentials
     /// when they expire, enabling long-running operations on remote storage.
     pub fn storage_options_provider(mut self, provider: Arc<dyn StorageOptionsProvider>) -> Self {
-        self.request
+        let store_params = self
+            .request
             .write_options
             .lance_write_params
             .get_or_insert(Default::default())
             .store_params
-            .get_or_insert(Default::default())
-            .storage_options_provider = Some(provider);
+            .get_or_insert(Default::default());
+        set_storage_options_provider(store_params, provider);
         self
     }
 }
@@ -450,15 +505,13 @@ impl OpenTableBuilder {
     ///
     /// See available options at <https://lancedb.com/docs/storage/>
     pub fn storage_option(mut self, key: impl Into<String>, value: impl Into<String>) -> Self {
-        let storage_options = self
+        let store_params = self
             .request
             .lance_read_params
             .get_or_insert(Default::default())
             .store_options
-            .get_or_insert(Default::default())
-            .storage_options
             .get_or_insert(Default::default());
-        storage_options.insert(key.into(), value.into());
+        merge_storage_options(store_params, [(key.into(), value.into())]);
         self
     }
 
@@ -472,18 +525,16 @@ impl OpenTableBuilder {
         mut self,
         pairs: impl IntoIterator<Item = (impl Into<String>, impl Into<String>)>,
     ) -> Self {
-        let storage_options = self
+        let store_params = self
             .request
             .lance_read_params
             .get_or_insert(Default::default())
             .store_options
-            .get_or_insert(Default::default())
-            .storage_options
             .get_or_insert(Default::default());
-
-        for (key, value) in pairs {
-            storage_options.insert(key.into(), value.into());
-        }
+        let updates = pairs
+            .into_iter()
+            .map(|(key, value)| (key.into(), value.into()));
+        merge_storage_options(store_params, updates);
         self
     }
 
@@ -507,12 +558,13 @@ impl OpenTableBuilder {
     /// This allows tables to automatically refresh cloud storage credentials
     /// when they expire, enabling long-running operations on remote storage.
     pub fn storage_options_provider(mut self, provider: Arc<dyn StorageOptionsProvider>) -> Self {
-        self.request
+        let store_params = self
+            .request
             .lance_read_params
             .get_or_insert(Default::default())
             .store_options
-            .get_or_insert(Default::default())
-            .storage_options_provider = Some(provider);
+            .get_or_insert(Default::default());
+        set_storage_options_provider(store_params, provider);
         self
     }
 
@@ -868,6 +920,10 @@ pub struct ConnectBuilder {
     embedding_registry: Option<Arc<dyn EmbeddingRegistry>>,
 }
 
+#[cfg(feature = "remote")]
+const ENV_VARS_TO_STORAGE_OPTS: [(&str, &str); 1] =
+    [("AZURE_STORAGE_ACCOUNT_NAME", "azure_storage_account_name")];
+
 impl ConnectBuilder {
     /// Create a new [`ConnectOptions`] with the given database URI.
     pub fn new(uri: &str) -> Self {
@@ -1051,11 +1107,27 @@ impl ConnectBuilder {
         self
     }
 
+    #[cfg(feature = "remote")]
+    fn apply_env_defaults(
+        env_var_to_remote_storage_option: &[(&str, &str)],
+        options: &mut HashMap<String, String>,
+    ) {
+        for (env_key, opt_key) in env_var_to_remote_storage_option {
+            if let Ok(env_value) = std::env::var(env_key) {
+                if !options.contains_key(*opt_key) {
+                    options.insert((*opt_key).to_string(), env_value);
+                }
+            }
+        }
+    }
+
     #[cfg(feature = "remote")]
     fn execute_remote(self) -> Result<Connection> {
         use crate::remote::db::RemoteDatabaseOptions;
 
-        let options = RemoteDatabaseOptions::parse_from_map(&self.request.options)?;
+        let mut merged_options = self.request.options.clone();
+        Self::apply_env_defaults(&ENV_VARS_TO_STORAGE_OPTS, &mut merged_options);
+        let options = RemoteDatabaseOptions::parse_from_map(&merged_options)?;
 
         let region = options.region.ok_or_else(|| Error::InvalidInput {
             message: "A region is required when connecting to LanceDb Cloud".to_string(),
@@ -1277,8 +1349,6 @@ mod test_utils {
 
 #[cfg(test)]
 mod tests {
-    use std::fs::create_dir_all;
-
     use crate::database::listing::{ListingDatabaseOptions, NewTableConfig};
     use crate::query::QueryBase;
     use crate::query::{ExecutableQuery, QueryExecutionOptions};
@@ -1302,6 +1372,23 @@ mod tests {
         assert_eq!(tc.connection.uri(), tc.uri);
     }
 
+    #[cfg(feature = "remote")]
+    #[test]
+    fn test_apply_env_defaults() {
+        let env_key = "TEST_APPLY_ENV_DEFAULTS_ENVIRONMENT_VARIABLE_ENV_KEY";
+        let env_val = "TEST_APPLY_ENV_DEFAULTS_ENVIRONMENT_VARIABLE_ENV_VAL";
+        let opts_key = "test_apply_env_defaults_environment_variable_opts_key";
+        std::env::set_var(env_key, env_val);
+
+        let mut options = HashMap::new();
+        ConnectBuilder::apply_env_defaults(&[(env_key, opts_key)], &mut options);
+        assert_eq!(Some(&env_val.to_string()), options.get(opts_key));
+
+        options.insert(opts_key.to_string(), "EXPLICIT-VALUE".to_string());
+        ConnectBuilder::apply_env_defaults(&[(env_key, opts_key)], &mut options);
+        assert_eq!(Some(&"EXPLICIT-VALUE".to_string()), options.get(opts_key));
+    }
+
     #[cfg(not(windows))]
     #[tokio::test]
     async fn test_connect_relative() {
@@ -1526,18 +1613,27 @@ mod tests {
 
     #[tokio::test]
     async fn drop_table() {
-        let tmp_dir = tempdir().unwrap();
+        let tc = new_test_connection().await.unwrap();
+        let db = tc.connection;
 
-        let uri = tmp_dir.path().to_str().unwrap();
-        let db = connect(uri).execute().await.unwrap();
+        if tc.is_remote {
+            // All the typical endpoints such as s3:///, file-object-store:///, etc. treat drop_table
+            // as idempotent.
+            assert!(db.drop_table("invalid_table", &[]).await.is_ok());
+        } else {
+            // The behavior of drop_table when using a file:/// endpoint differs from all other
+            // object providers, in that it returns an error when deleting a non-existent table.
+            assert!(matches!(
+                db.drop_table("invalid_table", &[]).await,
+                Err(crate::Error::TableNotFound { .. }),
+            ));
+        }
 
-        // drop non-exist table
-        assert!(matches!(
-            db.drop_table("invalid_table", &[]).await,
-            Err(crate::Error::TableNotFound { .. }),
-        ));
-
-        create_dir_all(tmp_dir.path().join("table1.lance")).unwrap();
+        let schema = Arc::new(Schema::new(vec![Field::new("x", DataType::Int32, false)]));
+        db.create_empty_table("table1", schema.clone())
+            .execute()
+            .await
+            .unwrap();
         db.drop_table("table1", &[]).await.unwrap();
 
         let tables = db.table_names().execute().await.unwrap();
@@ -1624,4 +1720,128 @@ mod tests {
         let cloned_count = cloned_table.count_rows(None).await.unwrap();
         assert_eq!(source_count, cloned_count);
     }
+
+    #[tokio::test]
+    async fn test_create_empty_table_with_embeddings() {
+        use crate::embeddings::{EmbeddingDefinition, EmbeddingFunction};
+        use arrow_array::{
+            Array, FixedSizeListArray, Float32Array, RecordBatch, RecordBatchIterator, StringArray,
+        };
+        use std::borrow::Cow;
+
+        #[derive(Debug, Clone)]
+        struct MockEmbedding {
+            dim: usize,
+        }
+
+        impl EmbeddingFunction for MockEmbedding {
+            fn name(&self) -> &str {
+                "test_embedding"
+            }
+
+            fn source_type(&self) -> Result<Cow<'_, DataType>> {
+                Ok(Cow::Owned(DataType::Utf8))
+            }
+
+            fn dest_type(&self) -> Result<Cow<'_, DataType>> {
+                Ok(Cow::Owned(DataType::new_fixed_size_list(
+                    DataType::Float32,
+                    self.dim as i32,
+                    true,
+                )))
+            }
+
+            fn compute_source_embeddings(&self, source: Arc<dyn Array>) -> Result<Arc<dyn Array>> {
+                let len = source.len();
+                let values = vec![1.0f32; len * self.dim];
+                let values = Arc::new(Float32Array::from(values));
+                let field = Arc::new(Field::new("item", DataType::Float32, true));
+                Ok(Arc::new(FixedSizeListArray::new(
+                    field,
+                    self.dim as i32,
+                    values,
+                    None,
+                )))
+            }
+
+            fn compute_query_embeddings(&self, _input: Arc<dyn Array>) -> Result<Arc<dyn Array>> {
+                unimplemented!()
+            }
+        }
+
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+        let db = connect(uri).execute().await.unwrap();
+
+        let embed_func = Arc::new(MockEmbedding { dim: 128 });
+        db.embedding_registry()
+            .register("test_embedding", embed_func.clone())
+            .unwrap();
+
+        let schema = Arc::new(Schema::new(vec![Field::new("name", DataType::Utf8, true)]));
+        let ed = EmbeddingDefinition {
+            source_column: "name".to_owned(),
+            dest_column: Some("name_embedding".to_owned()),
+            embedding_name: "test_embedding".to_owned(),
+        };
+
+        let table = db
+            .create_empty_table("test", schema)
+            .mode(CreateTableMode::Overwrite)
+            .add_embedding(ed)
+            .unwrap()
+            .execute()
+            .await
+            .unwrap();
+
+        let table_schema = table.schema().await.unwrap();
+        assert!(table_schema.column_with_name("name").is_some());
+        assert!(table_schema.column_with_name("name_embedding").is_some());
+
+        let embedding_field = table_schema.field_with_name("name_embedding").unwrap();
+        assert_eq!(
+            embedding_field.data_type(),
+            &DataType::new_fixed_size_list(DataType::Float32, 128, true)
+        );
+
+        let input_schema = Arc::new(Schema::new(vec![Field::new("name", DataType::Utf8, true)]));
+        let input_batch = RecordBatch::try_new(
+            input_schema.clone(),
+            vec![Arc::new(StringArray::from(vec![
+                Some("Alice"),
+                Some("Bob"),
+                Some("Charlie"),
+            ]))],
+        )
+        .unwrap();
+
+        let input_reader = Box::new(RecordBatchIterator::new(
+            vec![Ok(input_batch)].into_iter(),
+            input_schema,
+        ));
+
+        table.add(input_reader).execute().await.unwrap();
+
+        let results = table
+            .query()
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        assert_eq!(results.len(), 1);
+        let batch = &results[0];
+        assert_eq!(batch.num_rows(), 3);
+        assert!(batch.column_by_name("name_embedding").is_some());
+
+        let embedding_col = batch
+            .column_by_name("name_embedding")
+            .unwrap()
+            .as_any()
+            .downcast_ref::<FixedSizeListArray>()
+            .unwrap();
+        assert_eq!(embedding_col.len(), 3);
+    }
 }

rust/lancedb/src/database/listing.rs

@@ -12,7 +12,7 @@ use lance::dataset::{builder::DatasetBuilder, ReadParams, WriteMode};
 use lance::io::{ObjectStore, ObjectStoreParams, WrappingObjectStore};
 use lance_datafusion::utils::StreamingWriteSource;
 use lance_encoding::version::LanceFileVersion;
-use lance_io::object_store::StorageOptionsProvider;
+use lance_io::object_store::{StorageOptionsAccessor, StorageOptionsProvider};
 use lance_table::io::commit::commit_handler_from_url;
 use object_store::local::LocalFileSystem;
 use snafu::ResultExt;
@@ -356,7 +356,13 @@ impl ListingDatabase {
                     .clone()
                     .unwrap_or_else(|| Arc::new(lance::session::Session::default()));
                 let os_params = ObjectStoreParams {
-                    storage_options: Some(options.storage_options.clone()),
+                    storage_options_accessor: if options.storage_options.is_empty() {
+                        None
+                    } else {
+                        Some(Arc::new(StorageOptionsAccessor::with_static_options(
+                            options.storage_options.clone(),
+                        )))
+                    },
                     ..Default::default()
                 };
                 let (object_store, base_path) = ObjectStore::from_uri_and_params(
@@ -492,7 +498,13 @@ impl ListingDatabase {
 
     async fn drop_tables(&self, names: Vec<String>) -> Result<()> {
         let object_store_params = ObjectStoreParams {
-            storage_options: Some(self.storage_options.clone()),
+            storage_options_accessor: if self.storage_options.is_empty() {
+                None
+            } else {
+                Some(Arc::new(StorageOptionsAccessor::with_static_options(
+                    self.storage_options.clone(),
+                )))
+            },
             ..Default::default()
         };
         let mut uri = self.uri.clone();
@@ -541,7 +553,7 @@ impl ListingDatabase {
             .lance_write_params
             .as_ref()
             .and_then(|p| p.store_params.as_ref())
-            .and_then(|sp| sp.storage_options.as_ref());
+            .and_then(|sp| sp.storage_options());
 
         let storage_version_override = storage_options
             .and_then(|opts| opts.get(OPT_NEW_TABLE_STORAGE_VERSION))
@@ -592,21 +604,20 @@ impl ListingDatabase {
         // will cause a new connection to be created, and that connection will
         // be dropped from the cache when python GCs the table object, which
         // confounds reuse across tables.
-        if !self.storage_options.is_empty() {
-            let storage_options = write_params
+        if !self.storage_options.is_empty() || self.storage_options_provider.is_some() {
+            let store_params = write_params
                 .store_params
-                .get_or_insert_with(Default::default)
-                .storage_options
                 .get_or_insert_with(Default::default);
-            self.inherit_storage_options(storage_options);
-        }
-
-        // Set storage options provider if available
-        if self.storage_options_provider.is_some() {
-            write_params
-                .store_params
-                .get_or_insert_with(Default::default)
-                .storage_options_provider = self.storage_options_provider.clone();
+            let mut storage_options = store_params.storage_options().cloned().unwrap_or_default();
+            if !self.storage_options.is_empty() {
+                self.inherit_storage_options(&mut storage_options);
+            }
+            let accessor = if let Some(ref provider) = self.storage_options_provider {
+                StorageOptionsAccessor::with_initial_and_provider(storage_options, provider.clone())
+            } else {
+                StorageOptionsAccessor::with_static_options(storage_options)
+            };
+            store_params.storage_options_accessor = Some(Arc::new(accessor));
         }
 
         write_params.data_storage_version = self
@@ -892,7 +903,13 @@ impl Database for ListingDatabase {
         validate_table_name(&request.target_table_name)?;
 
         let storage_params = ObjectStoreParams {
-            storage_options: Some(self.storage_options.clone()),
+            storage_options_accessor: if self.storage_options.is_empty() {
+                None
+            } else {
+                Some(Arc::new(StorageOptionsAccessor::with_static_options(
+                    self.storage_options.clone(),
+                )))
+            },
             ..Default::default()
         };
         let read_params = ReadParams {
@@ -956,25 +973,28 @@ impl Database for ListingDatabase {
         // will cause a new connection to be created, and that connection will
         // be dropped from the cache when python GCs the table object, which
         // confounds reuse across tables.
-        if !self.storage_options.is_empty() {
-            let storage_options = request
+        if !self.storage_options.is_empty() || self.storage_options_provider.is_some() {
+            let store_params = request
                 .lance_read_params
                 .get_or_insert_with(Default::default)
                 .store_options
-                .get_or_insert_with(Default::default)
-                .storage_options
                 .get_or_insert_with(Default::default);
-            self.inherit_storage_options(storage_options);
-        }
-
-        // Set storage options provider if available
-        if self.storage_options_provider.is_some() {
-            request
-                .lance_read_params
-                .get_or_insert_with(Default::default)
-                .store_options
-                .get_or_insert_with(Default::default)
-                .storage_options_provider = self.storage_options_provider.clone();
+            let mut storage_options = store_params.storage_options().cloned().unwrap_or_default();
+            if !self.storage_options.is_empty() {
+                self.inherit_storage_options(&mut storage_options);
+            }
+            // Preserve request-level provider if no connection-level provider exists
+            let request_provider = store_params
+                .storage_options_accessor
+                .as_ref()
+                .and_then(|a| a.provider().cloned());
+            let provider = self.storage_options_provider.clone().or(request_provider);
+            let accessor = if let Some(provider) = provider {
+                StorageOptionsAccessor::with_initial_and_provider(storage_options, provider)
+            } else {
+                StorageOptionsAccessor::with_static_options(storage_options)
+            };
+            store_params.storage_options_accessor = Some(Arc::new(accessor));
         }
 
         // Some ReadParams are exposed in the OpenTableBuilder, but we also
@@ -1881,7 +1901,9 @@ mod tests {
         let write_options = WriteOptions {
             lance_write_params: Some(lance::dataset::WriteParams {
                 store_params: Some(lance::io::ObjectStoreParams {
-                    storage_options: Some(storage_options),
+                    storage_options_accessor: Some(Arc::new(
+                        StorageOptionsAccessor::with_static_options(storage_options),
+                    )),
                     ..Default::default()
                 }),
                 ..Default::default()
@@ -1955,7 +1977,9 @@ mod tests {
         let write_options = WriteOptions {
             lance_write_params: Some(lance::dataset::WriteParams {
                 store_params: Some(lance::io::ObjectStoreParams {
-                    storage_options: Some(storage_options),
+                    storage_options_accessor: Some(Arc::new(
+                        StorageOptionsAccessor::with_static_options(storage_options),
+                    )),
                     ..Default::default()
                 }),
                 ..Default::default()

rust/lancedb/src/database/namespace.rs

@@ -9,14 +9,15 @@ use std::sync::Arc;
 use async_trait::async_trait;
 use lance_namespace::{
     models::{
-        CreateNamespaceRequest, CreateNamespaceResponse, DeclareTableRequest,
-        DescribeNamespaceRequest, DescribeNamespaceResponse, DescribeTableRequest,
-        DropNamespaceRequest, DropNamespaceResponse, DropTableRequest, ListNamespacesRequest,
-        ListNamespacesResponse, ListTablesRequest, ListTablesResponse,
+        CreateEmptyTableRequest, CreateNamespaceRequest, CreateNamespaceResponse,
+        DeclareTableRequest, DescribeNamespaceRequest, DescribeNamespaceResponse,
+        DescribeTableRequest, DropNamespaceRequest, DropNamespaceResponse, DropTableRequest,
+        ListNamespacesRequest, ListNamespacesResponse, ListTablesRequest, ListTablesResponse,
     },
     LanceNamespace,
 };
 use lance_namespace_impls::ConnectBuilder;
+use log::warn;
 
 use crate::database::ReadConsistency;
 use crate::error::{Error, Result};
@@ -154,7 +155,6 @@ impl Database for LanceNamespaceDatabase {
         table_id.push(request.name.clone());
         let describe_request = DescribeTableRequest {
             id: Some(table_id.clone()),
-            version: None,
             ..Default::default()
         };
 
@@ -205,26 +205,53 @@ impl Database for LanceNamespaceDatabase {
         let mut table_id = request.namespace.clone();
         table_id.push(request.name.clone());
 
-        let create_empty_request = DeclareTableRequest {
+        // Try declare_table first, falling back to create_empty_table for backwards
+        // compatibility with older namespace clients that don't support declare_table
+        let declare_request = DeclareTableRequest {
             id: Some(table_id.clone()),
-            location: None,
-            vend_credentials: None,
             ..Default::default()
         };
 
-        let create_empty_response = self
-            .namespace
-            .declare_table(create_empty_request)
-            .await
-            .map_err(|e| Error::Runtime {
-                message: format!("Failed to declare table: {}", e),
-            })?;
+        let location = match self.namespace.declare_table(declare_request).await {
+            Ok(response) => response.location.ok_or_else(|| Error::Runtime {
+                message: "Table location is missing from declare_table response".to_string(),
+            })?,
+            Err(e) => {
+                // Check if the error is "not supported" and try create_empty_table as fallback
+                let err_str = e.to_string().to_lowercase();
+                if err_str.contains("not supported") || err_str.contains("not implemented") {
+                    warn!(
+                        "declare_table is not supported by the namespace client, \
+                        falling back to deprecated create_empty_table. \
+                        create_empty_table is deprecated and will be removed in Lance 3.0.0. \
+                        Please upgrade your namespace client to support declare_table."
+                    );
+                    #[allow(deprecated)]
+                    let create_empty_request = CreateEmptyTableRequest {
+                        id: Some(table_id.clone()),
+                        ..Default::default()
+                    };
 
-        let location = create_empty_response
-            .location
-            .ok_or_else(|| Error::Runtime {
-                message: "Table location is missing from create_empty_table response".to_string(),
-            })?;
+                    #[allow(deprecated)]
+                    let create_response = self
+                        .namespace
+                        .create_empty_table(create_empty_request)
+                        .await
+                        .map_err(|e| Error::Runtime {
+                            message: format!("Failed to create empty table: {}", e),
+                        })?;
+
+                    create_response.location.ok_or_else(|| Error::Runtime {
+                        message: "Table location is missing from create_empty_table response"
+                            .to_string(),
+                    })?
+                } else {
+                    return Err(Error::Runtime {
+                        message: format!("Failed to declare table: {}", e),
+                    });
+                }
+            }
+        };
 
         let native_table = NativeTable::create_from_namespace(
             self.namespace.clone(),
@@ -439,8 +466,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -501,8 +526,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -566,8 +589,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -651,8 +672,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -708,8 +727,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -790,8 +807,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await
@@ -825,8 +840,6 @@ mod tests {
         // Create a child namespace first
         conn.create_namespace(CreateNamespaceRequest {
             id: Some(vec!["test_ns".into()]),
-            mode: None,
-            properties: None,
             ..Default::default()
         })
         .await

rust/lancedb/src/dataloader/permutation/builder.rs

@@ -19,7 +19,7 @@ use crate::{
         split::{SplitStrategy, Splitter, SPLIT_ID_COLUMN},
         util::{rename_column, TemporaryDirectory},
     },
-    query::{ExecutableQuery, QueryBase},
+    query::{ExecutableQuery, QueryBase, Select},
     Error, Result, Table,
 };
 
@@ -27,6 +27,8 @@ pub const SRC_ROW_ID_COL: &str = "row_id";
 
 pub const SPLIT_NAMES_CONFIG_KEY: &str = "split_names";
 
+pub const DEFAULT_MEMORY_LIMIT: usize = 100 * 1024 * 1024;
+
 /// Where to store the permutation table
 #[derive(Debug, Clone, Default)]
 enum PermutationDestination {
@@ -167,10 +169,20 @@ impl PermutationBuilder {
         &self,
         data: SendableRecordBatchStream,
     ) -> Result<SendableRecordBatchStream> {
+        let memory_limit = std::env::var("LANCEDB_PERM_BUILDER_MEMORY_LIMIT")
+            .unwrap_or_else(|_| DEFAULT_MEMORY_LIMIT.to_string())
+            .parse::<usize>()
+            .unwrap_or_else(|_| {
+                log::error!(
+                    "Failed to parse LANCEDB_PERM_BUILDER_MEMORY_LIMIT, using default: {}",
+                    DEFAULT_MEMORY_LIMIT
+                );
+                DEFAULT_MEMORY_LIMIT
+            });
         let ctx = SessionContext::new_with_config_rt(
             SessionConfig::default(),
             RuntimeEnvBuilder::new()
-                .with_memory_limit(100 * 1024 * 1024, 1.0)
+                .with_memory_limit(memory_limit, 1.0)
                 .with_disk_manager_builder(
                     DiskManagerBuilder::default()
                         .with_mode(self.config.temp_dir.to_disk_manager_mode()),
@@ -232,7 +244,7 @@ impl PermutationBuilder {
     /// Builds the permutation table and stores it in the given database.
     pub async fn build(self) -> Result<Table> {
         // First pass, apply filter and load row ids
-        let mut rows = self.base_table.query().with_row_id();
+        let mut rows = self.base_table.query().select(Select::columns(&[ROW_ID]));
 
         if let Some(filter) = &self.config.filter {
             rows = rows.only_if(filter);
@@ -321,6 +333,47 @@ mod tests {
 
     use super::*;
 
+    #[tokio::test]
+    async fn test_permutation_table_only_stores_row_id_and_split_id() {
+        let temp_dir = tempfile::tempdir().unwrap();
+
+        let db = connect(temp_dir.path().to_str().unwrap())
+            .execute()
+            .await
+            .unwrap();
+
+        let initial_data = lance_datagen::gen_batch()
+            .col("col_a", lance_datagen::array::step::<Int32Type>())
+            .col("col_b", lance_datagen::array::step::<Int32Type>())
+            .into_ldb_stream(RowCount::from(100), BatchCount::from(10));
+        let data_table = db
+            .create_table_streaming("base_tbl", initial_data)
+            .execute()
+            .await
+            .unwrap();
+
+        let permutation_table = PermutationBuilder::new(data_table.clone())
+            .with_split_strategy(
+                SplitStrategy::Sequential {
+                    sizes: SplitSizes::Percentages(vec![0.5, 0.5]),
+                },
+                None,
+            )
+            .with_filter("col_a > 57".to_string())
+            .build()
+            .await
+            .unwrap();
+
+        let schema = permutation_table.schema().await.unwrap();
+        let field_names: Vec<&str> = schema.fields().iter().map(|f| f.name().as_str()).collect();
+        assert_eq!(
+            field_names,
+            vec!["row_id", "split_id"],
+            "Permutation table should only contain row_id and split_id columns, but found: {:?}",
+            field_names,
+        );
+    }
+
     #[tokio::test]
     async fn test_permutation_builder() {
         let temp_dir = tempfile::tempdir().unwrap();
@@ -352,8 +405,6 @@ mod tests {
             .await
             .unwrap();
 
-        println!("permutation_table: {:?}", permutation_table);
-
         // Potentially brittle seed-dependent values below
         assert_eq!(permutation_table.count_rows(None).await.unwrap(), 330);
         assert_eq!(

rust/lancedb/src/dataloader/permutation/shuffle.rs

@@ -171,7 +171,7 @@ impl Shuffler {
             // This is kind of an annoying limitation but if we allow runt clumps from batches then
             // clumps will get unaligned and we will mess up the clumps when we do the in-memory
             // shuffle step.  If this is a problem we can probably figure out a better way to do this.
-            if !is_last && batch.num_rows() as u64 % clump_size != 0 {
+            if !is_last && !(batch.num_rows() as u64).is_multiple_of(clump_size) {
                 return Err(Error::Runtime {
                     message: format!(
                         "Expected batch size ({}) to be divisible by clump size ({})",

rust/lancedb/src/dataloader/permutation/split.rs

@@ -1,12 +1,9 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
-use std::{
-    iter,
-    sync::{
-        atomic::{AtomicBool, AtomicU64, AtomicUsize, Ordering},
-        Arc,
-    },
+use std::sync::{
+    atomic::{AtomicBool, AtomicU64, AtomicUsize, Ordering},
+    Arc,
 };
 
 use arrow_array::{Array, BooleanArray, RecordBatch, UInt64Array};
@@ -15,6 +12,8 @@ use datafusion_common::hash_utils::create_hashes;
 use futures::{StreamExt, TryStreamExt};
 use lance_arrow::SchemaExt;
 
+use lance_core::ROW_ID;
+
 use crate::{
     arrow::{SendableRecordBatchStream, SimpleRecordBatchStream},
     dataloader::{
@@ -158,7 +157,7 @@ impl Splitter {
                 remaining_in_split
             };
 
-            split_ids.extend(iter::repeat(split_id as u64).take(rows_to_add as usize));
+            split_ids.extend(std::iter::repeat_n(split_id as u64, rows_to_add as usize));
             if done {
                 // Quit early if we've run out of splits
                 break;
@@ -363,11 +362,15 @@ impl Splitter {
 
     pub fn project(&self, query: Query) -> Query {
         match &self.strategy {
-            SplitStrategy::Calculated { calculation } => query.select(Select::Dynamic(vec![(
-                SPLIT_ID_COLUMN.to_string(),
-                calculation.clone(),
-            )])),
-            SplitStrategy::Hash { columns, .. } => query.select(Select::Columns(columns.clone())),
+            SplitStrategy::Calculated { calculation } => query.select(Select::Dynamic(vec![
+                (SPLIT_ID_COLUMN.to_string(), calculation.clone()),
+                (ROW_ID.to_string(), ROW_ID.to_string()),
+            ])),
+            SplitStrategy::Hash { columns, .. } => {
+                let mut cols = columns.clone();
+                cols.push(ROW_ID.to_string());
+                query.select(Select::Columns(cols))
+            }
             _ => query,
         }
     }
@@ -662,7 +665,7 @@ mod tests {
         assert_eq!(split_batch.num_rows(), total_split_sizes as usize);
         let mut expected = Vec::with_capacity(total_split_sizes as usize);
         for (i, size) in expected_split_sizes.iter().enumerate() {
-            expected.extend(iter::repeat(i as u64).take(*size as usize));
+            expected.extend(std::iter::repeat_n(i as u64, *size as usize));
         }
         let expected = Arc::new(UInt64Array::from(expected)) as Arc<dyn Array>;
 

rust/lancedb/src/index/vector.rs

@@ -297,10 +297,10 @@ impl IvfPqIndexBuilder {
 }
 
 pub(crate) fn suggested_num_sub_vectors(dim: u32) -> u32 {
-    if dim % 16 == 0 {
+    if dim.is_multiple_of(16) {
         // Should be more aggressive than this default.
         dim / 16
-    } else if dim % 8 == 0 {
+    } else if dim.is_multiple_of(8) {
         dim / 8
     } else {
         log::warn!(

rust/lancedb/src/lib.rs

@@ -51,24 +51,19 @@
 //! - `s3://bucket/path/to/database` or `gs://bucket/path/to/database` - database on cloud object store
 //! - `db://dbname` - Lance Cloud
 //!
-//! You can also use [`ConnectOptions`] to configure the connection to the database.
+//! You can also use [`ConnectBuilder`] to configure the connection to the database.
 //!
 //! ```rust
-//! # #[cfg(feature = "aws")]
-//! # {
-//! use object_store::aws::AwsCredential;
 //! # tokio::runtime::Runtime::new().unwrap().block_on(async {
 //! let db = lancedb::connect("data/sample-lancedb")
-//!     .aws_creds(AwsCredential {
-//!         key_id: "some_key".to_string(),
-//!         secret_key: "some_secret".to_string(),
-//!         token: None,
-//!     })
+//!     .storage_options([
+//!         ("aws_access_key_id", "some_key"),
+//!         ("aws_secret_access_key", "some_secret"),
+//!     ])
 //!     .execute()
 //!     .await
 //!     .unwrap();
 //! # });
-//! # }
 //! ```
 //!
 //! LanceDB uses [arrow-rs](https://github.com/apache/arrow-rs) to define schema, data types and array itself.

rust/lancedb/src/remote/db.rs

@@ -1718,8 +1718,6 @@ mod tests {
             let namespace = vec!["test_ns".to_string()];
             conn.create_namespace(CreateNamespaceRequest {
                 id: Some(namespace.clone()),
-                mode: None,
-                properties: None,
                 ..Default::default()
             })
             .await
@@ -1745,8 +1743,6 @@ mod tests {
             let list_response = conn
                 .list_tables(ListTablesRequest {
                     id: Some(namespace.clone()),
-                    page_token: None,
-                    limit: None,
                     ..Default::default()
                 })
                 .await
@@ -1758,8 +1754,6 @@ mod tests {
             let list_response = namespace_client
                 .list_tables(ListTablesRequest {
                     id: Some(namespace.clone()),
-                    page_token: None,
-                    limit: None,
                     ..Default::default()
                 })
                 .await
@@ -1800,8 +1794,6 @@ mod tests {
             let namespace = vec!["multi_table_ns".to_string()];
             conn.create_namespace(CreateNamespaceRequest {
                 id: Some(namespace.clone()),
-                mode: None,
-                properties: None,
                 ..Default::default()
             })
             .await
@@ -1827,8 +1819,6 @@ mod tests {
             let list_response = conn
                 .list_tables(ListTablesRequest {
                     id: Some(namespace.clone()),
-                    page_token: None,
-                    limit: None,
                     ..Default::default()
                 })
                 .await

rust/lancedb/src/remote/table.rs

@@ -1,6 +1,8 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
+pub mod insert;
+
 use crate::index::Index;
 use crate::index::IndexStatistics;
 use crate::query::{QueryFilter, QueryRequest, Select, VectorQueryRequest};
@@ -468,7 +470,9 @@ impl<S: HttpSend> RemoteTable<S> {
         self.apply_query_params(&mut body, &query.base)?;
 
         // Apply general parameters, before we dispatch based on number of query vectors.
-        body["distance_type"] = serde_json::json!(query.distance_type.unwrap_or_default());
+        if let Some(distance_type) = query.distance_type {
+            body["distance_type"] = serde_json::json!(distance_type);
+        }
         // In 0.23.1 we migrated from `nprobes` to `minimum_nprobes` and `maximum_nprobes`.
         // Old client / new server: since minimum_nprobes is missing, fallback to nprobes
         // New client / old server: old server will only see nprobes, make sure to set both
@@ -1493,6 +1497,14 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
         None
     }
 
+    async fn initial_storage_options(&self) -> Option<HashMap<String, String>> {
+        None
+    }
+
+    async fn latest_storage_options(&self) -> Result<Option<HashMap<String, String>>> {
+        Ok(None)
+    }
+
     async fn stats(&self) -> Result<TableStatistics> {
         let request = self
             .client
@@ -1508,6 +1520,21 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
         })?;
         Ok(stats)
     }
+
+    async fn create_insert_exec(
+        &self,
+        input: Arc<dyn ExecutionPlan>,
+        write_params: lance::dataset::WriteParams,
+    ) -> Result<Arc<dyn ExecutionPlan>> {
+        let overwrite = matches!(write_params.mode, lance::dataset::WriteMode::Overwrite);
+        Ok(Arc::new(insert::RemoteInsertExec::new(
+            self.name.clone(),
+            self.identifier.clone(),
+            self.client.clone(),
+            input,
+            overwrite,
+        )))
+    }
 }
 
 #[derive(Serialize)]
@@ -2230,7 +2257,6 @@ mod tests {
             let body: serde_json::Value = serde_json::from_slice(body).unwrap();
             let mut expected_body = serde_json::json!({
                 "prefilter": true,
-                "distance_type": "l2",
                 "nprobes": 20,
                 "minimum_nprobes": 20,
                 "maximum_nprobes": 20,

rust/lancedb/src/remote/table/insert.rs

@@ -0,0 +1,438 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! DataFusion ExecutionPlan for inserting data into remote LanceDB tables.
+
+use std::any::Any;
+use std::sync::{Arc, Mutex};
+
+use arrow_array::{ArrayRef, RecordBatch, UInt64Array};
+use arrow_ipc::CompressionType;
+use arrow_schema::ArrowError;
+use datafusion_common::{DataFusionError, Result as DataFusionResult};
+use datafusion_execution::{SendableRecordBatchStream, TaskContext};
+use datafusion_physical_expr::EquivalenceProperties;
+use datafusion_physical_plan::stream::RecordBatchStreamAdapter;
+use datafusion_physical_plan::{DisplayAs, DisplayFormatType, ExecutionPlan, PlanProperties};
+use futures::StreamExt;
+use http::header::CONTENT_TYPE;
+
+use crate::remote::client::{HttpSend, RestfulLanceDbClient, Sender};
+use crate::remote::table::RemoteTable;
+use crate::remote::ARROW_STREAM_CONTENT_TYPE;
+use crate::table::datafusion::insert::COUNT_SCHEMA;
+use crate::table::AddResult;
+use crate::Error;
+
+/// ExecutionPlan for inserting data into a remote LanceDB table.
+///
+/// This plan:
+/// 1. Requires single partition (no parallel remote inserts yet)
+/// 2. Streams data as Arrow IPC to `/v1/table/{id}/insert/` endpoint
+/// 3. Stores AddResult for retrieval after execution
+#[derive(Debug)]
+pub struct RemoteInsertExec<S: HttpSend = Sender> {
+    table_name: String,
+    identifier: String,
+    client: RestfulLanceDbClient<S>,
+    input: Arc<dyn ExecutionPlan>,
+    overwrite: bool,
+    properties: PlanProperties,
+    add_result: Arc<Mutex<Option<AddResult>>>,
+}
+
+impl<S: HttpSend + 'static> RemoteInsertExec<S> {
+    /// Create a new RemoteInsertExec.
+    pub fn new(
+        table_name: String,
+        identifier: String,
+        client: RestfulLanceDbClient<S>,
+        input: Arc<dyn ExecutionPlan>,
+        overwrite: bool,
+    ) -> Self {
+        let schema = COUNT_SCHEMA.clone();
+        let properties = PlanProperties::new(
+            EquivalenceProperties::new(schema),
+            datafusion_physical_plan::Partitioning::UnknownPartitioning(1),
+            datafusion_physical_plan::execution_plan::EmissionType::Final,
+            datafusion_physical_plan::execution_plan::Boundedness::Bounded,
+        );
+
+        Self {
+            table_name,
+            identifier,
+            client,
+            input,
+            overwrite,
+            properties,
+            add_result: Arc::new(Mutex::new(None)),
+        }
+    }
+
+    /// Get the add result after execution.
+    // TODO: this will be used when we wire this up to Table::add().
+    #[allow(dead_code)]
+    pub fn add_result(&self) -> Option<AddResult> {
+        self.add_result.lock().unwrap().clone()
+    }
+
+    fn stream_as_body(data: SendableRecordBatchStream) -> DataFusionResult<reqwest::Body> {
+        let options = arrow_ipc::writer::IpcWriteOptions::default()
+            .try_with_compression(Some(CompressionType::LZ4_FRAME))?;
+        let writer = arrow_ipc::writer::StreamWriter::try_new_with_options(
+            Vec::new(),
+            &data.schema(),
+            options,
+        )?;
+
+        let stream = futures::stream::try_unfold((data, writer), move |(mut data, mut writer)| {
+            async move {
+                match data.next().await {
+                    Some(Ok(batch)) => {
+                        writer.write(&batch)?;
+                        let buffer = std::mem::take(writer.get_mut());
+                        Ok(Some((buffer, (data, writer))))
+                    }
+                    Some(Err(e)) => Err(e),
+                    None => {
+                        if let Err(ArrowError::IpcError(_msg)) = writer.finish() {
+                            // Will error if already closed.
+                            return Ok(None);
+                        };
+                        let buffer = std::mem::take(writer.get_mut());
+                        Ok(Some((buffer, (data, writer))))
+                    }
+                }
+            }
+        });
+
+        Ok(reqwest::Body::wrap_stream(stream))
+    }
+}
+
+impl<S: HttpSend + 'static> DisplayAs for RemoteInsertExec<S> {
+    fn fmt_as(&self, t: DisplayFormatType, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match t {
+            DisplayFormatType::Default | DisplayFormatType::Verbose => {
+                write!(
+                    f,
+                    "RemoteInsertExec: table={}, overwrite={}",
+                    self.table_name, self.overwrite
+                )
+            }
+            DisplayFormatType::TreeRender => {
+                write!(f, "RemoteInsertExec")
+            }
+        }
+    }
+}
+
+impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
+    fn name(&self) -> &str {
+        Self::static_name()
+    }
+
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
+    fn properties(&self) -> &PlanProperties {
+        &self.properties
+    }
+
+    fn children(&self) -> Vec<&Arc<dyn ExecutionPlan>> {
+        vec![&self.input]
+    }
+
+    fn maintains_input_order(&self) -> Vec<bool> {
+        vec![false]
+    }
+
+    fn required_input_distribution(&self) -> Vec<datafusion_physical_plan::Distribution> {
+        // Until we have a separate commit endpoint, we need to do all inserts in a single partition
+        vec![datafusion_physical_plan::Distribution::SinglePartition]
+    }
+
+    fn benefits_from_input_partitioning(&self) -> Vec<bool> {
+        vec![false]
+    }
+
+    fn with_new_children(
+        self: Arc<Self>,
+        children: Vec<Arc<dyn ExecutionPlan>>,
+    ) -> DataFusionResult<Arc<dyn ExecutionPlan>> {
+        if children.len() != 1 {
+            return Err(DataFusionError::Internal(
+                "RemoteInsertExec requires exactly one child".to_string(),
+            ));
+        }
+        Ok(Arc::new(Self::new(
+            self.table_name.clone(),
+            self.identifier.clone(),
+            self.client.clone(),
+            children[0].clone(),
+            self.overwrite,
+        )))
+    }
+
+    fn execute(
+        &self,
+        partition: usize,
+        context: Arc<TaskContext>,
+    ) -> DataFusionResult<SendableRecordBatchStream> {
+        if partition != 0 {
+            return Err(DataFusionError::Internal(
+                "RemoteInsertExec only supports single partition execution".to_string(),
+            ));
+        }
+
+        let input_stream = self.input.execute(0, context)?;
+        let client = self.client.clone();
+        let identifier = self.identifier.clone();
+        let overwrite = self.overwrite;
+        let add_result = self.add_result.clone();
+        let table_name = self.table_name.clone();
+
+        let stream = futures::stream::once(async move {
+            let mut request = client
+                .post(&format!("/v1/table/{}/insert/", identifier))
+                .header(CONTENT_TYPE, ARROW_STREAM_CONTENT_TYPE);
+
+            if overwrite {
+                request = request.query(&[("mode", "overwrite")]);
+            }
+
+            let body = Self::stream_as_body(input_stream)?;
+            let request = request.body(body);
+
+            let (request_id, response) = client
+                .send(request)
+                .await
+                .map_err(|e| DataFusionError::External(Box::new(e)))?;
+
+            let response =
+                RemoteTable::<Sender>::handle_table_not_found(&table_name, response, &request_id)
+                    .await
+                    .map_err(|e| DataFusionError::External(Box::new(e)))?;
+
+            let response = client
+                .check_response(&request_id, response)
+                .await
+                .map_err(|e| DataFusionError::External(Box::new(e)))?;
+
+            let body_text = response.text().await.map_err(|e| {
+                DataFusionError::External(Box::new(Error::Http {
+                    source: Box::new(e),
+                    request_id: request_id.clone(),
+                    status_code: None,
+                }))
+            })?;
+
+            let parsed_result = if body_text.trim().is_empty() {
+                // Backward compatible with old servers
+                AddResult { version: 0 }
+            } else {
+                serde_json::from_str(&body_text).map_err(|e| {
+                    DataFusionError::External(Box::new(Error::Http {
+                        source: format!("Failed to parse add response: {}", e).into(),
+                        request_id: request_id.clone(),
+                        status_code: None,
+                    }))
+                })?
+            };
+
+            {
+                let mut res_lock = add_result.lock().map_err(|_| {
+                    DataFusionError::Execution("Failed to acquire lock for add_result".to_string())
+                })?;
+                *res_lock = Some(parsed_result);
+            }
+
+            // Return a single batch with count 0 (actual count is tracked in add_result)
+            let count_array: ArrayRef = Arc::new(UInt64Array::from(vec![0u64]));
+            let batch = RecordBatch::try_new(COUNT_SCHEMA.clone(), vec![count_array])?;
+            Ok::<_, DataFusionError>(batch)
+        });
+
+        Ok(Box::pin(RecordBatchStreamAdapter::new(
+            COUNT_SCHEMA.clone(),
+            stream,
+        )))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use arrow_array::record_batch;
+    use arrow_schema::{DataType, Field, Schema as ArrowSchema};
+    use datafusion::prelude::SessionContext;
+    use datafusion_catalog::MemTable;
+    use std::sync::atomic::{AtomicUsize, Ordering};
+    use std::sync::Arc;
+
+    use crate::remote::ARROW_STREAM_CONTENT_TYPE;
+    use crate::table::datafusion::BaseTableAdapter;
+    use crate::Table;
+
+    fn schema_json() -> &'static str {
+        r#"{"fields": [{"name": "id", "type": {"type": "int32"}, "nullable": true}]}"#
+    }
+
+    #[tokio::test]
+    async fn test_remote_insert_exec_execute_empty() {
+        let request_count = Arc::new(AtomicUsize::new(0));
+        let request_count_clone = request_count.clone();
+
+        let table = Table::new_with_handler("my_table", move |request| {
+            let path = request.url().path();
+
+            if path == "/v1/table/my_table/describe/" {
+                // Return schema for BaseTableAdapter::try_new
+                return http::Response::builder()
+                    .status(200)
+                    .body(format!(r#"{{"version": 1, "schema": {}}}"#, schema_json()))
+                    .unwrap();
+            }
+
+            if path == "/v1/table/my_table/insert/" {
+                assert_eq!(request.method(), "POST");
+                assert_eq!(
+                    request.headers().get("Content-Type").unwrap(),
+                    ARROW_STREAM_CONTENT_TYPE
+                );
+                request_count_clone.fetch_add(1, Ordering::SeqCst);
+
+                return http::Response::builder()
+                    .status(200)
+                    .body(r#"{"version": 2}"#.to_string())
+                    .unwrap();
+            }
+
+            panic!("Unexpected request path: {}", path);
+        });
+
+        let schema = Arc::new(ArrowSchema::new(vec![Field::new(
+            "id",
+            DataType::Int32,
+            true,
+        )]));
+
+        // Create empty MemTable (no batches)
+        let source_table = MemTable::try_new(schema, vec![vec![]]).unwrap();
+
+        let ctx = SessionContext::new();
+
+        // Register the remote table as insert target
+        let provider = BaseTableAdapter::try_new(table.base_table().clone())
+            .await
+            .unwrap();
+        ctx.register_table("my_table", Arc::new(provider)).unwrap();
+
+        // Register empty source
+        ctx.register_table("empty_source", Arc::new(source_table))
+            .unwrap();
+
+        // Execute the INSERT
+        ctx.sql("INSERT INTO my_table SELECT * FROM empty_source")
+            .await
+            .unwrap()
+            .collect()
+            .await
+            .unwrap();
+
+        // Verify: should have made exactly one HTTP request even with empty input
+        assert_eq!(request_count.load(Ordering::SeqCst), 1);
+    }
+
+    #[tokio::test]
+    async fn test_remote_insert_exec_multi_partition() {
+        let request_count = Arc::new(AtomicUsize::new(0));
+        let request_count_clone = request_count.clone();
+
+        let table = Table::new_with_handler("my_table", move |request| {
+            let path = request.url().path();
+
+            if path == "/v1/table/my_table/describe/" {
+                // Return schema for BaseTableAdapter::try_new
+                return http::Response::builder()
+                    .status(200)
+                    .body(format!(r#"{{"version": 1, "schema": {}}}"#, schema_json()))
+                    .unwrap();
+            }
+
+            if path == "/v1/table/my_table/insert/" {
+                assert_eq!(request.method(), "POST");
+                assert_eq!(
+                    request.headers().get("Content-Type").unwrap(),
+                    ARROW_STREAM_CONTENT_TYPE
+                );
+                request_count_clone.fetch_add(1, Ordering::SeqCst);
+
+                return http::Response::builder()
+                    .status(200)
+                    .body(r#"{"version": 2}"#.to_string())
+                    .unwrap();
+            }
+
+            panic!("Unexpected request path: {}", path);
+        });
+
+        let schema = Arc::new(ArrowSchema::new(vec![Field::new(
+            "id",
+            DataType::Int32,
+            true,
+        )]));
+
+        // Create MemTable with multiple partitions and multiple batches
+        let source_table = MemTable::try_new(
+            schema,
+            vec![
+                // Partition 0
+                vec![
+                    record_batch!(("id", Int32, [1, 2])).unwrap(),
+                    record_batch!(("id", Int32, [3, 4])).unwrap(),
+                ],
+                // Partition 1
+                vec![record_batch!(("id", Int32, [5, 6, 7])).unwrap()],
+                // Partition 2
+                vec![record_batch!(("id", Int32, [8])).unwrap()],
+            ],
+        )
+        .unwrap();
+
+        let ctx = SessionContext::new();
+
+        // Register the remote table as insert target
+        let provider = BaseTableAdapter::try_new(table.base_table().clone())
+            .await
+            .unwrap();
+        ctx.register_table("my_table", Arc::new(provider)).unwrap();
+
+        // Register multi-partition source
+        ctx.register_table("multi_partition_source", Arc::new(source_table))
+            .unwrap();
+
+        // Get the physical plan and verify it includes a repartition to 1
+        let df = ctx
+            .sql("INSERT INTO my_table SELECT * FROM multi_partition_source")
+            .await
+            .unwrap();
+        let plan = df.clone().create_physical_plan().await.unwrap();
+        let plan_str = datafusion::physical_plan::displayable(plan.as_ref())
+            .indent(true)
+            .to_string();
+
+        // The plan should include a CoalescePartitionsExec to merge partitions
+        assert!(
+            plan_str.contains("CoalescePartitionsExec"),
+            "Expected CoalescePartitionsExec in plan:\n{}",
+            plan_str
+        );
+
+        // Execute the INSERT
+        df.collect().await.unwrap();
+
+        // Verify: should have made exactly one HTTP request despite multiple input partitions
+        assert_eq!(request_count.load(Ordering::SeqCst), 1);
+    }
+}

rust/lancedb/src/table.rs

@@ -23,9 +23,7 @@ pub use lance::dataset::ColumnAlteration;
 pub use lance::dataset::NewColumnTransform;
 pub use lance::dataset::ReadParams;
 pub use lance::dataset::Version;
-use lance::dataset::{
-    InsertBuilder, UpdateBuilder as LanceUpdateBuilder, WhenMatched, WriteMode, WriteParams,
-};
+use lance::dataset::{InsertBuilder, WhenMatched, WriteMode, WriteParams};
 use lance::dataset::{MergeInsertBuilder as LanceMergeInsertBuilder, WhenNotMatchedBySource};
 use lance::index::vector::utils::infer_vector_dim;
 use lance::index::vector::VectorIndexParams;
@@ -40,7 +38,7 @@ use lance_index::vector::pq::PQBuildParams;
 use lance_index::vector::sq::builder::SQBuildParams;
 use lance_index::DatasetIndexExt;
 use lance_index::IndexType;
-use lance_io::object_store::LanceNamespaceStorageOptionsProvider;
+use lance_io::object_store::{LanceNamespaceStorageOptionsProvider, StorageOptionsAccessor};
 use lance_namespace::models::{
     QueryTableRequest as NsQueryTableRequest, QueryTableRequestColumns,
     QueryTableRequestFullTextQuery, QueryTableRequestVector, StringFtsQuery,
@@ -79,10 +77,14 @@ use self::merge::MergeInsertBuilder;
 
 pub mod datafusion;
 pub(crate) mod dataset;
+pub mod delete;
 pub mod merge;
+pub mod schema_evolution;
+pub mod update;
 
 use crate::index::waiter::wait_for_index;
 pub use chrono::Duration;
+pub use delete::DeleteResult;
 use futures::future::{join_all, Either};
 pub use lance::dataset::optimize::CompactionOptions;
 pub use lance::dataset::refs::{TagContents, Tags as LanceTags};
@@ -90,7 +92,9 @@ pub use lance::dataset::scanner::DatasetRecordBatchStream;
 use lance::dataset::statistics::DatasetStatisticsExt;
 use lance_index::frag_reuse::FRAG_REUSE_INDEX_NAME;
 pub use lance_index::optimize::OptimizeOptions;
+pub use schema_evolution::{AddColumnsResult, AlterColumnsResult, DropColumnsResult};
 use serde_with::skip_serializing_none;
+pub use update::{UpdateBuilder, UpdateResult};
 
 /// Defines the type of column
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -327,72 +331,6 @@ impl<T: IntoArrow> AddDataBuilder<T> {
     }
 }
 
-/// A builder for configuring an [`Table::update`] operation
-#[derive(Debug, Clone)]
-pub struct UpdateBuilder {
-    parent: Arc<dyn BaseTable>,
-    pub(crate) filter: Option<String>,
-    pub(crate) columns: Vec<(String, String)>,
-}
-
-impl UpdateBuilder {
-    fn new(parent: Arc<dyn BaseTable>) -> Self {
-        Self {
-            parent,
-            filter: None,
-            columns: Vec::new(),
-        }
-    }
-
-    /// Limits the update operation to rows matching the given filter
-    ///
-    /// If a row does not match the filter then it will be left unchanged.
-    pub fn only_if(mut self, filter: impl Into<String>) -> Self {
-        self.filter = Some(filter.into());
-        self
-    }
-
-    /// Specifies a column to update
-    ///
-    /// This method may be called multiple times to update multiple columns
-    ///
-    /// The `update_expr` should be an SQL expression explaining how to calculate
-    /// the new value for the column.  The expression will be evaluated against the
-    /// previous row's value.
-    ///
-    /// # Examples
-    ///
-    /// ```
-    /// # use lancedb::Table;
-    /// # async fn doctest_helper(tbl: Table) {
-    ///   let mut operation = tbl.update();
-    ///   // Increments the `bird_count` value by 1
-    ///   operation = operation.column("bird_count", "bird_count + 1");
-    ///   operation.execute().await.unwrap();
-    /// # }
-    /// ```
-    pub fn column(
-        mut self,
-        column_name: impl Into<String>,
-        update_expr: impl Into<String>,
-    ) -> Self {
-        self.columns.push((column_name.into(), update_expr.into()));
-        self
-    }
-
-    /// Executes the update operation.
-    /// Returns the update result
-    pub async fn execute(self) -> Result<UpdateResult> {
-        if self.columns.is_empty() {
-            Err(Error::InvalidInput {
-                message: "at least one column must be specified in an update operation".to_string(),
-            })
-        } else {
-            self.parent.clone().update(self).await
-        }
-    }
-}
-
 /// Filters that can be used to limit the rows returned by a query
 pub enum Filter {
     /// A SQL filter string
@@ -426,17 +364,6 @@ pub trait Tags: Send + Sync {
     async fn update(&mut self, tag: &str, version: u64) -> Result<()>;
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-pub struct UpdateResult {
-    #[serde(default)]
-    pub rows_updated: u64,
-    // The commit version associated with the operation.
-    // A version of `0` indicates compatibility with legacy servers that do not return
-    /// a commit version.
-    #[serde(default)]
-    pub version: u64,
-}
-
 #[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
 pub struct AddResult {
     // The commit version associated with the operation.
@@ -446,15 +373,6 @@ pub struct AddResult {
     pub version: u64,
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-pub struct DeleteResult {
-    // The commit version associated with the operation.
-    // A version of `0` indicates compatibility with legacy servers that do not return
-    /// a commit version.
-    #[serde(default)]
-    pub version: u64,
-}
-
 #[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
 pub struct MergeResult {
     // The commit version associated with the operation.
@@ -480,33 +398,6 @@ pub struct MergeResult {
     pub num_attempts: u32,
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-pub struct AddColumnsResult {
-    // The commit version associated with the operation.
-    // A version of `0` indicates compatibility with legacy servers that do not return
-    /// a commit version.
-    #[serde(default)]
-    pub version: u64,
-}
-
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-pub struct AlterColumnsResult {
-    // The commit version associated with the operation.
-    // A version of `0` indicates compatibility with legacy servers that do not return
-    /// a commit version.
-    #[serde(default)]
-    pub version: u64,
-}
-
-#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
-pub struct DropColumnsResult {
-    // The commit version associated with the operation.
-    // A version of `0` indicates compatibility with legacy servers that do not return
-    /// a commit version.
-    #[serde(default)]
-    pub version: u64,
-}
-
 /// A trait for anything "table-like".  This is used for both native tables (which target
 /// Lance datasets) and remote tables (which target LanceDB cloud)
 ///
@@ -611,7 +502,17 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
     /// Get the table URI (storage location)
     async fn uri(&self) -> Result<String>;
     /// Get the storage options used when opening this table, if any.
+    #[deprecated(since = "0.25.0", note = "Use initial_storage_options() instead")]
     async fn storage_options(&self) -> Option<HashMap<String, String>>;
+    /// Get the initial storage options that were passed in when opening this table.
+    ///
+    /// For dynamically refreshed options (e.g., credential vending), use [`Self::latest_storage_options`].
+    async fn initial_storage_options(&self) -> Option<HashMap<String, String>>;
+    /// Get the latest storage options, refreshing from provider if configured.
+    ///
+    /// Returns `Ok(Some(options))` if storage options are available (static or refreshed),
+    /// `Ok(None)` if no storage options were configured, or `Err(...)` if refresh failed.
+    async fn latest_storage_options(&self) -> Result<Option<HashMap<String, String>>>;
     /// Poll until the columns are fully indexed. Will return Error::Timeout if the columns
     /// are not fully indexed within the timeout.
     async fn wait_for_index(
@@ -621,6 +522,19 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
     ) -> Result<()>;
     /// Get statistics on the table
     async fn stats(&self) -> Result<TableStatistics>;
+    /// Create an ExecutionPlan for inserting data into the table.
+    ///
+    /// This is used by the DataFusion TableProvider implementation to support
+    /// INSERT INTO statements.
+    async fn create_insert_exec(
+        &self,
+        _input: Arc<dyn datafusion_physical_plan::ExecutionPlan>,
+        _write_params: WriteParams,
+    ) -> Result<Arc<dyn datafusion_physical_plan::ExecutionPlan>> {
+        Err(Error::NotSupported {
+            message: "create_insert_exec not implemented".to_string(),
+        })
+    }
 }
 
 /// A Table is a collection of strong typed Rows.
@@ -1328,10 +1242,32 @@ impl Table {
     /// Get the storage options used when opening this table, if any.
     ///
     /// Warning: This is an internal API and the return value is subject to change.
+    #[deprecated(since = "0.25.0", note = "Use initial_storage_options() instead")]
     pub async fn storage_options(&self) -> Option<HashMap<String, String>> {
+        #[allow(deprecated)]
         self.inner.storage_options().await
     }
 
+    /// Get the initial storage options that were passed in when opening this table.
+    ///
+    /// For dynamically refreshed options (e.g., credential vending), use [`Self::latest_storage_options`].
+    ///
+    /// Warning: This is an internal API and the return value is subject to change.
+    pub async fn initial_storage_options(&self) -> Option<HashMap<String, String>> {
+        self.inner.initial_storage_options().await
+    }
+
+    /// Get the latest storage options, refreshing from provider if configured.
+    ///
+    /// This method is useful for credential vending scenarios where storage options
+    /// may be refreshed dynamically. If no dynamic provider is configured, this
+    /// returns the initial static options.
+    ///
+    /// Warning: This is an internal API and the return value is subject to change.
+    pub async fn latest_storage_options(&self) -> Result<Option<HashMap<String, String>>> {
+        self.inner.latest_storage_options().await
+    }
+
     /// Get statistics about an index.
     /// Returns None if the index does not exist.
     pub async fn index_stats(
@@ -1425,8 +1361,8 @@ impl Table {
             })
             .collect::<Vec<_>>();
 
-        let unioned = UnionExec::try_new(projected_plans).map_err(|e| Error::Runtime {
-            message: format!("Failed to build union plan: {e}"),
+        let unioned = UnionExec::try_new(projected_plans).map_err(|err| Error::Runtime {
+            message: err.to_string(),
         })?;
         // We require 1 partition in the final output
         let repartitioned = RepartitionExec::try_new(
@@ -1668,18 +1604,14 @@ impl NativeTable {
 
         // Use DatasetBuilder::from_namespace which automatically fetches location
         // and storage options from the namespace
-        let builder = DatasetBuilder::from_namespace(
-            namespace_client.clone(),
-            table_id,
-            false, // Don't ignore namespace storage options
-        )
-        .await
-        .map_err(|e| match e {
-            lance::Error::Namespace { source, .. } => Error::Runtime {
-                message: format!("Failed to get table info from namespace: {:?}", source),
-            },
-            source => Error::Lance { source },
-        })?;
+        let builder = DatasetBuilder::from_namespace(namespace_client.clone(), table_id)
+            .await
+            .map_err(|e| match e {
+                lance::Error::Namespace { source, .. } => Error::Runtime {
+                    message: format!("Failed to get table info from namespace: {:?}", source),
+                },
+                source => Error::Lance { source },
+            })?;
 
         let dataset = builder
             .with_read_params(params)
@@ -1883,7 +1815,13 @@ impl NativeTable {
         let store_params = params
             .store_params
             .get_or_insert_with(ObjectStoreParams::default);
-        store_params.storage_options_provider = Some(storage_options_provider);
+        let accessor = match store_params.storage_options().cloned() {
+            Some(options) => {
+                StorageOptionsAccessor::with_initial_and_provider(options, storage_options_provider)
+            }
+            None => StorageOptionsAccessor::with_provider(storage_options_provider),
+        };
+        store_params.storage_options_accessor = Some(Arc::new(accessor));
 
         // Patch the params if we have a write store wrapper
         let params = match write_store_wrapper.clone() {
@@ -2059,7 +1997,7 @@ impl NativeTable {
             return provided;
         }
         let suggested = suggested_num_sub_vectors(dim);
-        if num_bits.is_some_and(|num_bits| num_bits == 4) && suggested % 2 != 0 {
+        if num_bits.is_some_and(|num_bits| num_bits == 4) && !suggested.is_multiple_of(2) {
             // num_sub_vectors must be even when 4 bits are used
             suggested + 1
         } else {
@@ -2349,7 +2287,7 @@ impl NativeTable {
                 };
 
                 // Convert select to columns list
-                let columns: Option<Box<QueryTableRequestColumns>> = match &vq.base.select {
+                let columns = match &vq.base.select {
                     Select::All => None,
                     Select::Columns(cols) => Some(Box::new(QueryTableRequestColumns {
                         column_names: Some(cols.clone()),
@@ -2407,7 +2345,6 @@ impl NativeTable {
                     with_row_id: Some(vq.base.with_row_id),
                     bypass_vector_index: Some(!vq.use_index),
                     full_text_query,
-                    version: None,
                     ..Default::default()
                 })
             }
@@ -2426,7 +2363,7 @@ impl NativeTable {
                     .map(|f| self.filter_to_sql(f))
                     .transpose()?;
 
-                let columns: Option<Box<QueryTableRequestColumns>> = match &q.select {
+                let columns = match &q.select {
                     Select::All => None,
                     Select::Columns(cols) => Some(Box::new(QueryTableRequestColumns {
                         column_names: Some(cols.clone()),
@@ -2470,18 +2407,10 @@ impl NativeTable {
                     columns,
                     prefilter: Some(q.prefilter),
                     offset: q.offset.map(|o| o as i32),
-                    ef: None,
-                    refine_factor: None,
-                    distance_type: None,
-                    nprobes: None,
                     vector_column: None, // No vector column for plain queries
                     with_row_id: Some(q.with_row_id),
                     bypass_vector_index: Some(true), // No vector index for plain queries
                     full_text_query,
-                    version: None,
-                    fast_search: None,
-                    lower_bound: None,
-                    upper_bound: None,
                     ..Default::default()
                 })
             }
@@ -2811,25 +2740,8 @@ impl BaseTable for NativeTable {
     }
 
     async fn update(&self, update: UpdateBuilder) -> Result<UpdateResult> {
-        let dataset = self.dataset.get().await?.clone();
-        let mut builder = LanceUpdateBuilder::new(Arc::new(dataset));
-        if let Some(predicate) = update.filter {
-            builder = builder.update_where(&predicate)?;
-        }
-
-        for (column, value) in update.columns {
-            builder = builder.set(column, &value)?;
-        }
-
-        let operation = builder.build()?;
-        let res = operation.execute().await?;
-        self.dataset
-            .set_latest(res.new_dataset.as_ref().clone())
-            .await;
-        Ok(UpdateResult {
-            rows_updated: res.rows_updated,
-            version: res.new_dataset.version().version,
-        })
+        // Delegate to the submodule implementation
+        update::execute_update(self, update).await
     }
 
     async fn create_plan(
@@ -3087,11 +2999,8 @@ impl BaseTable for NativeTable {
 
     /// Delete rows from the table
     async fn delete(&self, predicate: &str) -> Result<DeleteResult> {
-        let mut dataset = self.dataset.get_mut().await?;
-        dataset.delete(predicate).await?;
-        Ok(DeleteResult {
-            version: dataset.version().version,
-        })
+        // Delegate to the submodule implementation
+        delete::execute_delete(self, predicate).await
     }
 
     async fn tags(&self) -> Result<Box<dyn Tags + '_>> {
@@ -3157,27 +3066,15 @@ impl BaseTable for NativeTable {
         transforms: NewColumnTransform,
         read_columns: Option<Vec<String>>,
     ) -> Result<AddColumnsResult> {
-        let mut dataset = self.dataset.get_mut().await?;
-        dataset.add_columns(transforms, read_columns, None).await?;
-        Ok(AddColumnsResult {
-            version: dataset.version().version,
-        })
+        schema_evolution::execute_add_columns(self, transforms, read_columns).await
     }
 
     async fn alter_columns(&self, alterations: &[ColumnAlteration]) -> Result<AlterColumnsResult> {
-        let mut dataset = self.dataset.get_mut().await?;
-        dataset.alter_columns(alterations).await?;
-        Ok(AlterColumnsResult {
-            version: dataset.version().version,
-        })
+        schema_evolution::execute_alter_columns(self, alterations).await
     }
 
     async fn drop_columns(&self, columns: &[&str]) -> Result<DropColumnsResult> {
-        let mut dataset = self.dataset.get_mut().await?;
-        dataset.drop_columns(columns).await?;
-        Ok(DropColumnsResult {
-            version: dataset.version().version,
-        })
+        schema_evolution::execute_drop_columns(self, columns).await
     }
 
     async fn list_indices(&self) -> Result<Vec<IndexConfig>> {
@@ -3240,11 +3137,20 @@ impl BaseTable for NativeTable {
     }
 
     async fn storage_options(&self) -> Option<HashMap<String, String>> {
+        self.initial_storage_options().await
+    }
+
+    async fn initial_storage_options(&self) -> Option<HashMap<String, String>> {
         self.dataset
             .get()
             .await
             .ok()
-            .and_then(|dataset| dataset.storage_options().cloned())
+            .and_then(|dataset| dataset.initial_storage_options().cloned())
+    }
+
+    async fn latest_storage_options(&self) -> Result<Option<HashMap<String, String>>> {
+        let dataset = self.dataset.get().await?;
+        Ok(dataset.latest_storage_options().await?.map(|o| o.0))
     }
 
     async fn index_stats(&self, index_name: &str) -> Result<Option<IndexStatistics>> {
@@ -3360,6 +3266,21 @@ impl BaseTable for NativeTable {
         };
         Ok(stats)
     }
+
+    async fn create_insert_exec(
+        &self,
+        input: Arc<dyn datafusion_physical_plan::ExecutionPlan>,
+        write_params: WriteParams,
+    ) -> Result<Arc<dyn datafusion_physical_plan::ExecutionPlan>> {
+        let ds = self.dataset.get().await?;
+        let dataset = Arc::new((*ds).clone());
+        Ok(Arc::new(datafusion::insert::InsertExec::new(
+            self.dataset.clone(),
+            dataset,
+            input,
+            write_params,
+        )))
+    }
 }
 
 #[skip_serializing_none]
@@ -3409,22 +3330,18 @@ pub struct FragmentSummaryStats {
 #[cfg(test)]
 #[allow(deprecated)]
 mod tests {
-    use std::iter;
     use std::sync::atomic::{AtomicBool, Ordering};
     use std::sync::Arc;
     use std::time::Duration;
 
     use arrow_array::{
         builder::{ListBuilder, StringBuilder},
-        Array, BooleanArray, Date32Array, FixedSizeListArray, Float32Array, Float64Array,
-        Int32Array, Int64Array, LargeStringArray, RecordBatch, RecordBatchIterator,
-        RecordBatchReader, StringArray, TimestampMillisecondArray, TimestampNanosecondArray,
-        UInt32Array,
+        Array, BooleanArray, FixedSizeListArray, Float32Array, Int32Array, LargeStringArray,
+        RecordBatch, RecordBatchIterator, RecordBatchReader, StringArray,
     };
     use arrow_array::{BinaryArray, LargeBinaryArray};
     use arrow_data::ArrayDataBuilder;
-    use arrow_schema::{DataType, Field, Schema, TimeUnit};
-    use futures::TryStreamExt;
+    use arrow_schema::{DataType, Field, Schema};
     use lance::dataset::WriteMode;
     use lance::io::{ObjectStoreParams, WrappingObjectStore};
     use lance::Dataset;
@@ -3436,7 +3353,6 @@ mod tests {
     use crate::connection::ConnectBuilder;
     use crate::index::scalar::{BTreeIndexBuilder, BitmapIndexBuilder};
     use crate::index::vector::{IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder};
-    use crate::query::{ExecutableQuery, QueryBase};
 
     #[tokio::test]
     async fn test_open() {
@@ -3658,306 +3574,6 @@ mod tests {
         assert_eq!(table.name(), "test");
     }
 
-    #[tokio::test]
-    async fn test_update_with_predicate() {
-        let tmp_dir = tempdir().unwrap();
-        let dataset_path = tmp_dir.path().join("test.lance");
-        let uri = dataset_path.to_str().unwrap();
-        let conn = connect(uri)
-            .read_consistency_interval(Duration::from_secs(0))
-            .execute()
-            .await
-            .unwrap();
-
-        let schema = Arc::new(Schema::new(vec![
-            Field::new("id", DataType::Int32, false),
-            Field::new("name", DataType::Utf8, false),
-        ]));
-
-        let record_batch_iter = RecordBatchIterator::new(
-            vec![RecordBatch::try_new(
-                schema.clone(),
-                vec![
-                    Arc::new(Int32Array::from_iter_values(0..10)),
-                    Arc::new(StringArray::from_iter_values(vec![
-                        "a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
-                    ])),
-                ],
-            )
-            .unwrap()]
-            .into_iter()
-            .map(Ok),
-            schema.clone(),
-        );
-
-        let table = conn
-            .create_table("my_table", record_batch_iter)
-            .execute()
-            .await
-            .unwrap();
-
-        table
-            .update()
-            .only_if("id > 5")
-            .column("name", "'foo'")
-            .execute()
-            .await
-            .unwrap();
-
-        let mut batches = table
-            .query()
-            .select(Select::columns(&["id", "name"]))
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-
-        while let Some(batch) = batches.pop() {
-            let ids = batch
-                .column(0)
-                .as_any()
-                .downcast_ref::<Int32Array>()
-                .unwrap()
-                .iter()
-                .collect::<Vec<_>>();
-            let names = batch
-                .column(1)
-                .as_any()
-                .downcast_ref::<StringArray>()
-                .unwrap()
-                .iter()
-                .collect::<Vec<_>>();
-            for (i, name) in names.iter().enumerate() {
-                let id = ids[i].unwrap();
-                let name = name.unwrap();
-                if id > 5 {
-                    assert_eq!(name, "foo");
-                } else {
-                    assert_eq!(name, &format!("{}", (b'a' + id as u8) as char));
-                }
-            }
-        }
-    }
-
-    #[tokio::test]
-    async fn test_update_all_types() {
-        let tmp_dir = tempdir().unwrap();
-        let dataset_path = tmp_dir.path().join("test.lance");
-        let uri = dataset_path.to_str().unwrap();
-        let conn = connect(uri)
-            .read_consistency_interval(Duration::from_secs(0))
-            .execute()
-            .await
-            .unwrap();
-
-        let schema = Arc::new(Schema::new(vec![
-            Field::new("int32", DataType::Int32, false),
-            Field::new("int64", DataType::Int64, false),
-            Field::new("uint32", DataType::UInt32, false),
-            Field::new("string", DataType::Utf8, false),
-            Field::new("large_string", DataType::LargeUtf8, false),
-            Field::new("float32", DataType::Float32, false),
-            Field::new("float64", DataType::Float64, false),
-            Field::new("bool", DataType::Boolean, false),
-            Field::new("date32", DataType::Date32, false),
-            Field::new(
-                "timestamp_ns",
-                DataType::Timestamp(TimeUnit::Nanosecond, None),
-                false,
-            ),
-            Field::new(
-                "timestamp_ms",
-                DataType::Timestamp(TimeUnit::Millisecond, None),
-                false,
-            ),
-            Field::new(
-                "vec_f32",
-                DataType::FixedSizeList(Arc::new(Field::new("item", DataType::Float32, true)), 2),
-                false,
-            ),
-            Field::new(
-                "vec_f64",
-                DataType::FixedSizeList(Arc::new(Field::new("item", DataType::Float64, true)), 2),
-                false,
-            ),
-        ]));
-
-        let record_batch_iter = RecordBatchIterator::new(
-            vec![RecordBatch::try_new(
-                schema.clone(),
-                vec![
-                    Arc::new(Int32Array::from_iter_values(0..10)),
-                    Arc::new(Int64Array::from_iter_values(0..10)),
-                    Arc::new(UInt32Array::from_iter_values(0..10)),
-                    Arc::new(StringArray::from_iter_values(vec![
-                        "a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
-                    ])),
-                    Arc::new(LargeStringArray::from_iter_values(vec![
-                        "a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
-                    ])),
-                    Arc::new(Float32Array::from_iter_values((0..10).map(|i| i as f32))),
-                    Arc::new(Float64Array::from_iter_values((0..10).map(|i| i as f64))),
-                    Arc::new(Into::<BooleanArray>::into(vec![
-                        true, false, true, false, true, false, true, false, true, false,
-                    ])),
-                    Arc::new(Date32Array::from_iter_values(0..10)),
-                    Arc::new(TimestampNanosecondArray::from_iter_values(0..10)),
-                    Arc::new(TimestampMillisecondArray::from_iter_values(0..10)),
-                    Arc::new(
-                        create_fixed_size_list(
-                            Float32Array::from_iter_values((0..20).map(|i| i as f32)),
-                            2,
-                        )
-                        .unwrap(),
-                    ),
-                    Arc::new(
-                        create_fixed_size_list(
-                            Float64Array::from_iter_values((0..20).map(|i| i as f64)),
-                            2,
-                        )
-                        .unwrap(),
-                    ),
-                ],
-            )
-            .unwrap()]
-            .into_iter()
-            .map(Ok),
-            schema.clone(),
-        );
-
-        let table = conn
-            .create_table("my_table", record_batch_iter)
-            .execute()
-            .await
-            .unwrap();
-
-        // check it can do update for each type
-        let updates: Vec<(&str, &str)> = vec![
-            ("string", "'foo'"),
-            ("large_string", "'large_foo'"),
-            ("int32", "1"),
-            ("int64", "1"),
-            ("uint32", "1"),
-            ("float32", "1.0"),
-            ("float64", "1.0"),
-            ("bool", "true"),
-            ("date32", "1"),
-            ("timestamp_ns", "1"),
-            ("timestamp_ms", "1"),
-            ("vec_f32", "[1.0, 1.0]"),
-            ("vec_f64", "[1.0, 1.0]"),
-        ];
-
-        let mut update_op = table.update();
-        for (column, value) in updates {
-            update_op = update_op.column(column, value);
-        }
-        update_op.execute().await.unwrap();
-
-        let mut batches = table
-            .query()
-            .select(Select::columns(&[
-                "string",
-                "large_string",
-                "int32",
-                "int64",
-                "uint32",
-                "float32",
-                "float64",
-                "bool",
-                "date32",
-                "timestamp_ns",
-                "timestamp_ms",
-                "vec_f32",
-                "vec_f64",
-            ]))
-            .execute()
-            .await
-            .unwrap()
-            .try_collect::<Vec<_>>()
-            .await
-            .unwrap();
-        let batch = batches.pop().unwrap();
-
-        macro_rules! assert_column {
-            ($column:expr, $array_type:ty, $expected:expr) => {
-                let array = $column
-                    .as_any()
-                    .downcast_ref::<$array_type>()
-                    .unwrap()
-                    .iter()
-                    .collect::<Vec<_>>();
-                for v in array {
-                    assert_eq!(v, Some($expected));
-                }
-            };
-        }
-
-        assert_column!(batch.column(0), StringArray, "foo");
-        assert_column!(batch.column(1), LargeStringArray, "large_foo");
-        assert_column!(batch.column(2), Int32Array, 1);
-        assert_column!(batch.column(3), Int64Array, 1);
-        assert_column!(batch.column(4), UInt32Array, 1);
-        assert_column!(batch.column(5), Float32Array, 1.0);
-        assert_column!(batch.column(6), Float64Array, 1.0);
-        assert_column!(batch.column(7), BooleanArray, true);
-        assert_column!(batch.column(8), Date32Array, 1);
-        assert_column!(batch.column(9), TimestampNanosecondArray, 1);
-        assert_column!(batch.column(10), TimestampMillisecondArray, 1);
-
-        let array = batch
-            .column(11)
-            .as_any()
-            .downcast_ref::<FixedSizeListArray>()
-            .unwrap()
-            .iter()
-            .collect::<Vec<_>>();
-        for v in array {
-            let v = v.unwrap();
-            let f32array = v.as_any().downcast_ref::<Float32Array>().unwrap();
-            for v in f32array {
-                assert_eq!(v, Some(1.0));
-            }
-        }
-
-        let array = batch
-            .column(12)
-            .as_any()
-            .downcast_ref::<FixedSizeListArray>()
-            .unwrap()
-            .iter()
-            .collect::<Vec<_>>();
-        for v in array {
-            let v = v.unwrap();
-            let f64array = v.as_any().downcast_ref::<Float64Array>().unwrap();
-            for v in f64array {
-                assert_eq!(v, Some(1.0));
-            }
-        }
-    }
-
-    #[tokio::test]
-    async fn test_update_via_expr() {
-        let tmp_dir = tempdir().unwrap();
-        let dataset_path = tmp_dir.path().join("test.lance");
-        let uri = dataset_path.to_str().unwrap();
-        let conn = connect(uri)
-            .read_consistency_interval(Duration::from_secs(0))
-            .execute()
-            .await
-            .unwrap();
-        let tbl = conn
-            .create_table("my_table", make_test_batches())
-            .execute()
-            .await
-            .unwrap();
-        assert_eq!(1, tbl.count_rows(Some("i == 0".to_string())).await.unwrap());
-        tbl.update().column("i", "i+1").execute().await.unwrap();
-        assert_eq!(0, tbl.count_rows(Some("i == 0".to_string())).await.unwrap());
-    }
-
     #[derive(Default, Debug)]
     struct NoOpCacheWrapper {
         called: AtomicBool,
@@ -4026,7 +3642,7 @@ mod tests {
                 schema.clone(),
                 vec![
                     Arc::new(Int32Array::from_iter_values(offset..(offset + 10))),
-                    Arc::new(Int32Array::from_iter_values(iter::repeat(age).take(10))),
+                    Arc::new(Int32Array::from_iter_values(std::iter::repeat_n(age, 10))),
                 ],
             )],
             schema,
@@ -5154,15 +4770,16 @@ mod tests {
         let any_query = AnyQuery::VectorQuery(vq);
         let ns_request = table.convert_to_namespace_query(&any_query).unwrap();
 
-        let column_names = ns_request
-            .columns
-            .as_ref()
-            .and_then(|cols| cols.column_names.clone());
-
         assert_eq!(ns_request.k, 10);
         assert_eq!(ns_request.offset, Some(5));
         assert_eq!(ns_request.filter, Some("id > 0".to_string()));
-        assert_eq!(column_names, Some(vec!["id".to_string()]));
+        assert_eq!(
+            ns_request
+                .columns
+                .as_ref()
+                .and_then(|c| c.column_names.as_ref()),
+            Some(&vec!["id".to_string()])
+        );
         assert_eq!(ns_request.vector_column, Some("vector".to_string()));
         assert_eq!(ns_request.distance_type, Some("l2".to_string()));
         assert!(ns_request.vector.single_vector.is_some());
@@ -5199,16 +4816,17 @@ mod tests {
         let any_query = AnyQuery::Query(q);
         let ns_request = table.convert_to_namespace_query(&any_query).unwrap();
 
-        let column_names = ns_request
-            .columns
-            .as_ref()
-            .and_then(|cols| cols.column_names.clone());
-
         // Plain queries should pass an empty vector
         assert_eq!(ns_request.k, 20);
         assert_eq!(ns_request.offset, Some(5));
         assert_eq!(ns_request.filter, Some("id > 5".to_string()));
-        assert_eq!(column_names, Some(vec!["id".to_string()]));
+        assert_eq!(
+            ns_request
+                .columns
+                .as_ref()
+                .and_then(|c| c.column_names.as_ref()),
+            Some(&vec!["id".to_string()])
+        );
         assert_eq!(ns_request.with_row_id, Some(true));
         assert_eq!(ns_request.bypass_vector_index, Some(true));
         assert!(ns_request.vector_column.is_none()); // No vector column for plain queries

rust/lancedb/src/table/datafusion.rs

@@ -3,6 +3,7 @@
 
 //! This module contains adapters to allow LanceDB tables to be used as DataFusion table providers.
 
+pub mod insert;
 pub mod udtf;
 
 use std::{collections::HashMap, sync::Arc};
@@ -13,11 +14,12 @@ use async_trait::async_trait;
 use datafusion_catalog::{Session, TableProvider};
 use datafusion_common::{DataFusionError, Result as DataFusionResult, Statistics};
 use datafusion_execution::{SendableRecordBatchStream, TaskContext};
-use datafusion_expr::{Expr, TableProviderFilterPushDown, TableType};
+use datafusion_expr::{dml::InsertOp, Expr, TableProviderFilterPushDown, TableType};
 use datafusion_physical_plan::{
     stream::RecordBatchStreamAdapter, DisplayAs, DisplayFormatType, ExecutionPlan, PlanProperties,
 };
 use futures::{TryFutureExt, TryStreamExt};
+use lance::dataset::{WriteMode, WriteParams};
 
 use super::{AnyQuery, BaseTable};
 use crate::{
@@ -250,6 +252,33 @@ impl TableProvider for BaseTableAdapter {
         // TODO
         None
     }
+
+    async fn insert_into(
+        &self,
+        _state: &dyn Session,
+        input: Arc<dyn ExecutionPlan>,
+        insert_op: InsertOp,
+    ) -> DataFusionResult<Arc<dyn ExecutionPlan>> {
+        let mode = match insert_op {
+            InsertOp::Append => WriteMode::Append,
+            InsertOp::Overwrite => WriteMode::Overwrite,
+            InsertOp::Replace => {
+                return Err(DataFusionError::NotImplemented(
+                    "Replace mode is not supported for LanceDB tables".to_string(),
+                ))
+            }
+        };
+
+        let write_params = WriteParams {
+            mode,
+            ..Default::default()
+        };
+
+        self.table
+            .create_insert_exec(input, write_params)
+            .await
+            .map_err(|e| DataFusionError::External(e.into()))
+    }
 }
 
 #[cfg(test)]

rust/lancedb/src/table/datafusion/insert.rs

@@ -0,0 +1,446 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! DataFusion ExecutionPlan for inserting data into LanceDB tables.
+
+use std::any::Any;
+use std::sync::{Arc, LazyLock, Mutex};
+
+use arrow_array::{RecordBatch, UInt64Array};
+use arrow_schema::{DataType, Field, Schema as ArrowSchema, SchemaRef};
+use datafusion_common::{DataFusionError, Result as DataFusionResult};
+use datafusion_execution::{SendableRecordBatchStream, TaskContext};
+use datafusion_physical_expr::{EquivalenceProperties, Partitioning};
+use datafusion_physical_plan::execution_plan::{Boundedness, EmissionType};
+use datafusion_physical_plan::stream::RecordBatchStreamAdapter;
+use datafusion_physical_plan::{
+    DisplayAs, DisplayFormatType, ExecutionPlan, ExecutionPlanProperties, PlanProperties,
+};
+use lance::dataset::transaction::{Operation, Transaction};
+use lance::dataset::{CommitBuilder, InsertBuilder, WriteParams};
+use lance::Dataset;
+use lance_table::format::Fragment;
+
+use crate::table::dataset::DatasetConsistencyWrapper;
+
+pub(crate) static COUNT_SCHEMA: LazyLock<SchemaRef> = LazyLock::new(|| {
+    Arc::new(ArrowSchema::new(vec![Field::new(
+        "count",
+        DataType::UInt64,
+        false,
+    )]))
+});
+
+fn operation_fragments(operation: &Operation) -> &[Fragment] {
+    match operation {
+        Operation::Append { fragments } => fragments,
+        Operation::Overwrite { fragments, .. } => fragments,
+        _ => &[],
+    }
+}
+
+fn count_rows_from_operation(operation: &Operation) -> u64 {
+    operation_fragments(operation)
+        .iter()
+        .map(|f| f.num_rows().unwrap_or(0) as u64)
+        .sum()
+}
+
+fn operation_fragments_mut(operation: &mut Operation) -> &mut Vec<Fragment> {
+    match operation {
+        Operation::Append { fragments } => fragments,
+        Operation::Overwrite { fragments, .. } => fragments,
+        _ => panic!("Unsupported operation type for getting mutable fragments"),
+    }
+}
+
+fn merge_transactions(mut transactions: Vec<Transaction>) -> Option<Transaction> {
+    let mut first = transactions.pop()?;
+
+    for txn in transactions {
+        let first_fragments = operation_fragments_mut(&mut first.operation);
+        let txn_fragments = operation_fragments(&txn.operation);
+        first_fragments.extend_from_slice(txn_fragments);
+    }
+
+    Some(first)
+}
+
+/// ExecutionPlan for inserting data into a native LanceDB table.
+///
+/// This plan executes inserts by:
+/// 1. Each partition writes data independently using InsertBuilder::execute_uncommitted_stream
+/// 2. The last partition to complete commits all transactions atomically
+/// 3. Returns the count of inserted rows per partition
+#[derive(Debug)]
+pub struct InsertExec {
+    ds_wrapper: DatasetConsistencyWrapper,
+    dataset: Arc<Dataset>,
+    input: Arc<dyn ExecutionPlan>,
+    write_params: WriteParams,
+    properties: PlanProperties,
+    partial_transactions: Arc<Mutex<Vec<Transaction>>>,
+}
+
+impl InsertExec {
+    pub fn new(
+        ds_wrapper: DatasetConsistencyWrapper,
+        dataset: Arc<Dataset>,
+        input: Arc<dyn ExecutionPlan>,
+        write_params: WriteParams,
+    ) -> Self {
+        let schema = COUNT_SCHEMA.clone();
+        let num_partitions = input.output_partitioning().partition_count();
+        let properties = PlanProperties::new(
+            EquivalenceProperties::new(schema),
+            Partitioning::UnknownPartitioning(num_partitions),
+            EmissionType::Final,
+            Boundedness::Bounded,
+        );
+
+        Self {
+            ds_wrapper,
+            dataset,
+            input,
+            write_params,
+            properties,
+            partial_transactions: Arc::new(Mutex::new(Vec::with_capacity(num_partitions))),
+        }
+    }
+}
+
+impl DisplayAs for InsertExec {
+    fn fmt_as(&self, t: DisplayFormatType, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match t {
+            DisplayFormatType::Default | DisplayFormatType::Verbose => {
+                write!(f, "InsertExec: mode={:?}", self.write_params.mode)
+            }
+            DisplayFormatType::TreeRender => {
+                write!(f, "InsertExec")
+            }
+        }
+    }
+}
+
+impl ExecutionPlan for InsertExec {
+    fn name(&self) -> &str {
+        Self::static_name()
+    }
+
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
+    fn properties(&self) -> &PlanProperties {
+        &self.properties
+    }
+
+    fn children(&self) -> Vec<&Arc<dyn ExecutionPlan>> {
+        vec![&self.input]
+    }
+
+    fn maintains_input_order(&self) -> Vec<bool> {
+        vec![false]
+    }
+
+    fn benefits_from_input_partitioning(&self) -> Vec<bool> {
+        vec![false]
+    }
+
+    fn with_new_children(
+        self: Arc<Self>,
+        children: Vec<Arc<dyn ExecutionPlan>>,
+    ) -> DataFusionResult<Arc<dyn ExecutionPlan>> {
+        if children.len() != 1 {
+            return Err(DataFusionError::Internal(
+                "InsertExec requires exactly one child".to_string(),
+            ));
+        }
+        Ok(Arc::new(Self::new(
+            self.ds_wrapper.clone(),
+            self.dataset.clone(),
+            children[0].clone(),
+            self.write_params.clone(),
+        )))
+    }
+
+    fn execute(
+        &self,
+        partition: usize,
+        context: Arc<TaskContext>,
+    ) -> DataFusionResult<SendableRecordBatchStream> {
+        let input_stream = self.input.execute(partition, context)?;
+        let dataset = self.dataset.clone();
+        let write_params = self.write_params.clone();
+        let partial_transactions = self.partial_transactions.clone();
+        let total_partitions = self.input.output_partitioning().partition_count();
+        let ds_wrapper = self.ds_wrapper.clone();
+
+        let stream = futures::stream::once(async move {
+            let transaction = InsertBuilder::new(dataset.clone())
+                .with_params(&write_params)
+                .execute_uncommitted_stream(input_stream)
+                .await?;
+
+            let num_rows = count_rows_from_operation(&transaction.operation);
+
+            let to_commit = {
+                // Don't hold the lock over an await point.
+                let mut txns = partial_transactions.lock().unwrap();
+                txns.push(transaction);
+                if txns.len() == total_partitions {
+                    Some(std::mem::take(&mut *txns))
+                } else {
+                    None
+                }
+            };
+
+            if let Some(transactions) = to_commit {
+                if let Some(merged_txn) = merge_transactions(transactions) {
+                    let new_dataset = CommitBuilder::new(dataset.clone())
+                        .execute(merged_txn)
+                        .await?;
+                    ds_wrapper.set_latest(new_dataset).await;
+                }
+            }
+
+            Ok(RecordBatch::try_new(
+                COUNT_SCHEMA.clone(),
+                vec![Arc::new(UInt64Array::from(vec![num_rows]))],
+            )?)
+        });
+
+        Ok(Box::pin(RecordBatchStreamAdapter::new(
+            COUNT_SCHEMA.clone(),
+            stream,
+        )))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use std::vec;
+
+    use super::*;
+    use arrow_array::{record_batch, Int32Array, RecordBatchIterator};
+    use datafusion::prelude::SessionContext;
+    use datafusion_catalog::MemTable;
+    use tempfile::tempdir;
+
+    use crate::connect;
+
+    #[tokio::test]
+    async fn test_insert_via_sql() {
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+
+        let db = connect(uri).execute().await.unwrap();
+
+        // Create initial table
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+        let reader = RecordBatchIterator::new(vec![Ok(batch)], schema);
+
+        let table = db
+            .create_table("test_insert", Box::new(reader))
+            .execute()
+            .await
+            .unwrap();
+
+        // Verify initial count
+        assert_eq!(table.count_rows(None).await.unwrap(), 3);
+
+        let ctx = SessionContext::new();
+        let provider =
+            crate::table::datafusion::BaseTableAdapter::try_new(table.base_table().clone())
+                .await
+                .unwrap();
+        ctx.register_table("test_insert", Arc::new(provider))
+            .unwrap();
+
+        ctx.sql("INSERT INTO test_insert VALUES (4), (5), (6)")
+            .await
+            .unwrap()
+            .collect()
+            .await
+            .unwrap();
+
+        // Verify final count
+        table.checkout_latest().await.unwrap();
+        assert_eq!(table.count_rows(None).await.unwrap(), 6);
+    }
+
+    #[tokio::test]
+    async fn test_insert_overwrite_via_sql() {
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+
+        let db = connect(uri).execute().await.unwrap();
+
+        // Create initial table with 3 rows
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+        let reader = RecordBatchIterator::new(vec![Ok(batch)], schema);
+
+        let table = db
+            .create_table("test_overwrite", Box::new(reader))
+            .execute()
+            .await
+            .unwrap();
+
+        assert_eq!(table.count_rows(None).await.unwrap(), 3);
+
+        let ctx = SessionContext::new();
+        let provider =
+            crate::table::datafusion::BaseTableAdapter::try_new(table.base_table().clone())
+                .await
+                .unwrap();
+        ctx.register_table("test_overwrite", Arc::new(provider))
+            .unwrap();
+
+        ctx.sql("INSERT OVERWRITE INTO test_overwrite VALUES (10), (20)")
+            .await
+            .unwrap()
+            .collect()
+            .await
+            .unwrap();
+
+        // Verify: should have 2 rows (overwritten, not appended)
+        table.checkout_latest().await.unwrap();
+        assert_eq!(table.count_rows(None).await.unwrap(), 2);
+    }
+
+    #[tokio::test]
+    async fn test_insert_empty_batch() {
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+
+        let db = connect(uri).execute().await.unwrap();
+
+        // Create initial table
+        let schema = Arc::new(ArrowSchema::new(vec![Field::new(
+            "id",
+            DataType::Int32,
+            false,
+        )]));
+        let batches = vec![RecordBatch::try_new(
+            schema.clone(),
+            vec![Arc::new(Int32Array::from(vec![1, 2, 3]))],
+        )
+        .unwrap()];
+        let reader = RecordBatchIterator::new(batches.into_iter().map(Ok), schema.clone());
+
+        let table = db
+            .create_table("test_empty", Box::new(reader))
+            .execute()
+            .await
+            .unwrap();
+
+        assert_eq!(table.count_rows(None).await.unwrap(), 3);
+
+        let ctx = SessionContext::new();
+        let provider =
+            crate::table::datafusion::BaseTableAdapter::try_new(table.base_table().clone())
+                .await
+                .unwrap();
+        ctx.register_table("test_empty", Arc::new(provider))
+            .unwrap();
+
+        let source_schema = Arc::new(ArrowSchema::new(vec![Field::new(
+            "id",
+            DataType::Int32,
+            false,
+        )]));
+        // Empty batches
+        let source_reader = RecordBatchIterator::new(
+            std::iter::empty::<Result<RecordBatch, arrow_schema::ArrowError>>(),
+            source_schema,
+        );
+        let source_table = db
+            .create_table("empty_source", Box::new(source_reader))
+            .execute()
+            .await
+            .unwrap();
+        let source_provider =
+            crate::table::datafusion::BaseTableAdapter::try_new(source_table.base_table().clone())
+                .await
+                .unwrap();
+        ctx.register_table("empty_source", Arc::new(source_provider))
+            .unwrap();
+
+        // Execute INSERT with empty source
+        ctx.sql("INSERT INTO test_empty SELECT * FROM empty_source")
+            .await
+            .unwrap()
+            .collect()
+            .await
+            .unwrap();
+
+        // Verify: should still have 3 rows (nothing inserted)
+        table.checkout_latest().await.unwrap();
+        assert_eq!(table.count_rows(None).await.unwrap(), 3);
+    }
+
+    #[tokio::test]
+    async fn test_insert_multiple_batches() {
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+
+        let db = connect(uri).execute().await.unwrap();
+
+        // Create initial table
+        let schema = Arc::new(ArrowSchema::new(vec![Field::new(
+            "id",
+            DataType::Int32,
+            true,
+        )]));
+        let batches =
+            vec![
+                RecordBatch::try_new(schema.clone(), vec![Arc::new(Int32Array::from(vec![1]))])
+                    .unwrap(),
+            ];
+        let reader = RecordBatchIterator::new(batches.into_iter().map(Ok), schema.clone());
+
+        let table = db
+            .create_table("test_multi_batch", Box::new(reader))
+            .execute()
+            .await
+            .unwrap();
+
+        let ctx = SessionContext::new();
+        let provider =
+            crate::table::datafusion::BaseTableAdapter::try_new(table.base_table().clone())
+                .await
+                .unwrap();
+        ctx.register_table("test_multi_batch", Arc::new(provider))
+            .unwrap();
+
+        // Memtable with multiple batches and multiple partitions
+        let source_table = MemTable::try_new(
+            schema.clone(),
+            vec![
+                // Partition 0
+                vec![
+                    record_batch!(("id", Int32, [2, 3])).unwrap(),
+                    record_batch!(("id", Int32, [4, 5])).unwrap(),
+                ],
+                // Partition 1
+                vec![record_batch!(("id", Int32, [6, 7, 8])).unwrap()],
+            ],
+        )
+        .unwrap();
+        ctx.register_table("multi_batch_source", Arc::new(source_table))
+            .unwrap();
+
+        ctx.sql("INSERT INTO test_multi_batch SELECT * FROM multi_batch_source")
+            .await
+            .unwrap()
+            .collect()
+            .await
+            .unwrap();
+
+        // Verify: should have 1 + 2 + 2 + 3 = 8 rows
+        table.checkout_latest().await.unwrap();
+        assert_eq!(table.count_rows(None).await.unwrap(), 8);
+    }
+}

rust/lancedb/src/table/delete.rs

@@ -0,0 +1,161 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+use serde::{Deserialize, Serialize};
+
+use super::NativeTable;
+use crate::Result;
+
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
+pub struct DeleteResult {
+    // The commit version associated with the operation.
+    // A version of `0` indicates compatibility with legacy servers that do not return
+    /// a commit version.
+    #[serde(default)]
+    pub version: u64,
+}
+
+/// Internal implementation of the delete logic
+///
+/// This logic was moved from NativeTable::delete to keep table.rs clean.
+pub(crate) async fn execute_delete(table: &NativeTable, predicate: &str) -> Result<DeleteResult> {
+    // We access the dataset from the table. Since this is in the same module hierarchy (super),
+    // and 'dataset' is pub(crate), we can access it.
+    let mut dataset = table.dataset.get_mut().await?;
+
+    // Perform the actual delete on the Lance dataset
+    dataset.delete(predicate).await?;
+
+    // Return the result with the new version
+    Ok(DeleteResult {
+        version: dataset.version().version,
+    })
+}
+
+#[cfg(test)]
+mod tests {
+    use crate::connect;
+    use arrow_array::{record_batch, Int32Array, RecordBatch, RecordBatchIterator};
+    use arrow_schema::{DataType, Field, Schema};
+    use std::sync::Arc;
+
+    use crate::query::ExecutableQuery;
+    use futures::TryStreamExt;
+    #[tokio::test]
+    async fn test_delete_simple() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        // 1. Create a table with values 0 to 9
+        let schema = Arc::new(Schema::new(vec![Field::new("i", DataType::Int32, false)]));
+        let batch = RecordBatch::try_new(
+            schema.clone(),
+            vec![Arc::new(Int32Array::from_iter_values(0..10))],
+        )
+        .unwrap();
+
+        let table = conn
+            .create_table(
+                "test_delete",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // 2. Verify initial state
+        assert_eq!(table.count_rows(None).await.unwrap(), 10);
+
+        // 3. Execute Delete (removes values > 5)
+        table.delete("i > 5").await.unwrap();
+
+        // 4. Verify results
+        assert_eq!(table.count_rows(None).await.unwrap(), 6); // 0, 1, 2, 3, 4, 5 remain
+
+        // 5. Verify specific data consistency
+        let batches = table
+            .query()
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+        let batch = &batches[0];
+        let array = batch
+            .column(0)
+            .as_any()
+            .downcast_ref::<Int32Array>()
+            .unwrap();
+
+        // Ensure no value > 5 exists
+        for val in array.iter() {
+            assert!(val.unwrap() <= 5);
+        }
+    }
+    #[tokio::test]
+    async fn rows_removed_schema_same() {
+        let conn = connect("memory://").execute().await.unwrap();
+        let batch = record_batch!(
+            ("id", Int32, [1, 2, 3, 4, 5]),
+            ("name", Utf8, ["a", "b", "c", "d", "e"])
+        )
+        .unwrap();
+        let original_schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_delete_all",
+                RecordBatchIterator::new(vec![Ok(batch)], original_schema.clone()),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        table.delete("true").await.unwrap();
+
+        assert_eq!(table.count_rows(None).await.unwrap(), 0);
+
+        let current_schema = table.schema().await.unwrap();
+        //check if the original schema is the same as current
+        assert_eq!(current_schema, original_schema);
+    }
+
+    #[tokio::test]
+    async fn test_delete_false_increments_version() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        // Create a table with 5 rows
+        let batch = record_batch!(("id", Int32, [1, 2, 3, 4, 5])).unwrap();
+
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_delete_noop",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Capture the initial state (Rows = 5, Version = 1)
+        let initial_rows = table.count_rows(None).await.unwrap();
+        let initial_version = table.version().await.unwrap();
+
+        assert_eq!(initial_rows, 5);
+        table.delete("false").await.unwrap();
+
+        // Rows should still be 5
+        let current_rows = table.count_rows(None).await.unwrap();
+        assert_eq!(
+            current_rows, initial_rows,
+            "Data should not change when predicate is false"
+        );
+
+        // version check
+        let current_version = table.version().await.unwrap();
+        assert!(
+            current_version > initial_version,
+            "Table version must increment after delete operation"
+        );
+    }
+}

rust/lancedb/src/table/schema_evolution.rs

@@ -0,0 +1,666 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! Schema evolution operations for LanceDB tables.
+//!
+//! This module provides functionality to modify the schema of existing tables:
+//! - [`add_columns`](execute_add_columns): Add new columns using SQL expressions
+//! - [`alter_columns`](execute_alter_columns): Rename columns, change types, or modify nullability
+//! - [`drop_columns`](execute_drop_columns): Remove columns from the table
+
+use lance::dataset::{ColumnAlteration, NewColumnTransform};
+use serde::{Deserialize, Serialize};
+
+use super::NativeTable;
+use crate::Result;
+
+/// The result of an add columns operation.
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
+pub struct AddColumnsResult {
+    // The commit version associated with the operation.
+    // A version of `0` indicates compatibility with legacy servers that do not return
+    /// a commit version.
+    #[serde(default)]
+    pub version: u64,
+}
+
+/// The result of an alter columns operation.
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
+pub struct AlterColumnsResult {
+    // The commit version associated with the operation.
+    // A version of `0` indicates compatibility with legacy servers that do not return
+    /// a commit version.
+    #[serde(default)]
+    pub version: u64,
+}
+
+/// The result of a drop columns operation.
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
+pub struct DropColumnsResult {
+    // The commit version associated with the operation.
+    // A version of `0` indicates compatibility with legacy servers that do not return
+    /// a commit version.
+    #[serde(default)]
+    pub version: u64,
+}
+
+/// Internal implementation of the add columns logic.
+///
+/// Adds new columns to the table using the provided transforms.
+pub(crate) async fn execute_add_columns(
+    table: &NativeTable,
+    transforms: NewColumnTransform,
+    read_columns: Option<Vec<String>>,
+) -> Result<AddColumnsResult> {
+    let mut dataset = table.dataset.get_mut().await?;
+    dataset.add_columns(transforms, read_columns, None).await?;
+    Ok(AddColumnsResult {
+        version: dataset.version().version,
+    })
+}
+
+/// Internal implementation of the alter columns logic.
+///
+/// Alters existing columns in the table (rename, change type, or modify nullability).
+pub(crate) async fn execute_alter_columns(
+    table: &NativeTable,
+    alterations: &[ColumnAlteration],
+) -> Result<AlterColumnsResult> {
+    let mut dataset = table.dataset.get_mut().await?;
+    dataset.alter_columns(alterations).await?;
+    Ok(AlterColumnsResult {
+        version: dataset.version().version,
+    })
+}
+
+/// Internal implementation of the drop columns logic.
+///
+/// Removes columns from the table.
+pub(crate) async fn execute_drop_columns(
+    table: &NativeTable,
+    columns: &[&str],
+) -> Result<DropColumnsResult> {
+    let mut dataset = table.dataset.get_mut().await?;
+    dataset.drop_columns(columns).await?;
+    Ok(DropColumnsResult {
+        version: dataset.version().version,
+    })
+}
+
+#[cfg(test)]
+mod tests {
+    use arrow_array::{record_batch, Int32Array, RecordBatchIterator, StringArray};
+    use arrow_schema::DataType;
+    use futures::TryStreamExt;
+    use lance::dataset::ColumnAlteration;
+
+    use crate::connect;
+    use crate::query::{ExecutableQuery, QueryBase, Select};
+    use crate::table::NewColumnTransform;
+
+    // Add Columns Tests
+
+    #[tokio::test]
+    async fn test_add_columns_with_sql_expression() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("id", Int32, [1, 2, 3, 4, 5])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_add_columns",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        let initial_version = table.version().await.unwrap();
+
+        // Add a computed column
+        let result = table
+            .add_columns(
+                NewColumnTransform::SqlExpressions(vec![("doubled".into(), "id * 2".into())]),
+                None,
+            )
+            .await
+            .unwrap();
+
+        // Version should increment
+        assert!(result.version > initial_version);
+
+        // Verify the new column exists with correct values
+        let batches = table
+            .query()
+            .select(Select::columns(&["id", "doubled"]))
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        let batch = &batches[0];
+        let ids: Vec<i32> = batch
+            .column(0)
+            .as_any()
+            .downcast_ref::<Int32Array>()
+            .unwrap()
+            .iter()
+            .map(|v| v.unwrap())
+            .collect();
+        let doubled: Vec<i32> = batch
+            .column(1)
+            .as_any()
+            .downcast_ref::<Int32Array>()
+            .unwrap()
+            .iter()
+            .map(|v| v.unwrap())
+            .collect();
+
+        for (id, d) in ids.iter().zip(doubled.iter()) {
+            assert_eq!(*d, id * 2);
+        }
+    }
+
+    #[tokio::test]
+    async fn test_add_multiple_columns() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("x", Int32, [10, 20, 30])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_add_multi_columns",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Add multiple columns at once
+        table
+            .add_columns(
+                NewColumnTransform::SqlExpressions(vec![
+                    ("y".into(), "x + 1".into()),
+                    ("z".into(), "x * x".into()),
+                ]),
+                None,
+            )
+            .await
+            .unwrap();
+
+        // Verify schema has all columns
+        let schema = table.schema().await.unwrap();
+        assert_eq!(schema.fields().len(), 3);
+        assert!(schema.field_with_name("x").is_ok());
+        assert!(schema.field_with_name("y").is_ok());
+        assert!(schema.field_with_name("z").is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_add_column_with_constant_expression() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_add_const_column",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Add a column with a constant value
+        table
+            .add_columns(
+                NewColumnTransform::SqlExpressions(vec![("constant".into(), "42".into())]),
+                None,
+            )
+            .await
+            .unwrap();
+
+        let schema = table.schema().await.unwrap();
+        assert!(schema.field_with_name("constant").is_ok());
+
+        // Verify all values are 42
+        let batches = table
+            .query()
+            .select(Select::columns(&["constant"]))
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        let batch = &batches[0];
+        let values = batch["constant"]
+            .as_any()
+            .downcast_ref::<arrow_array::Int64Array>()
+            .unwrap()
+            .values();
+        assert!(values.iter().all(|&v| v == 42));
+    }
+
+    // Alter Columns Tests
+
+    #[tokio::test]
+    async fn test_alter_column_rename() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("old_name", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_alter_rename",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        let initial_version = table.version().await.unwrap();
+
+        // Rename the column
+        let result = table
+            .alter_columns(&[ColumnAlteration::new("old_name".into()).rename("new_name".into())])
+            .await
+            .unwrap();
+
+        // Version should increment
+        assert!(result.version > initial_version);
+
+        // Verify rename
+        let schema = table.schema().await.unwrap();
+        assert!(schema.field_with_name("old_name").is_err());
+        assert!(schema.field_with_name("new_name").is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_alter_column_set_nullable() {
+        use arrow_array::RecordBatch;
+        use arrow_schema::{Field, Schema};
+        use std::sync::Arc;
+
+        let conn = connect("memory://").execute().await.unwrap();
+
+        // Create a schema with a non-nullable field
+        let schema = Arc::new(Schema::new(vec![Field::new(
+            "value",
+            DataType::Int32,
+            false,
+        )]));
+        let batch = RecordBatch::try_new(
+            schema.clone(),
+            vec![Arc::new(Int32Array::from(vec![1, 2, 3]))],
+        )
+        .unwrap();
+
+        let table = conn
+            .create_table(
+                "test_alter_nullable",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Initially non-nullable
+        let schema = table.schema().await.unwrap();
+        assert!(!schema.field_with_name("value").unwrap().is_nullable());
+
+        // Make it nullable
+        table
+            .alter_columns(&[ColumnAlteration::new("value".into()).set_nullable(true)])
+            .await
+            .unwrap();
+
+        // Verify it's now nullable
+        let schema = table.schema().await.unwrap();
+        assert!(schema.field_with_name("value").unwrap().is_nullable());
+    }
+
+    #[tokio::test]
+    async fn test_alter_column_cast_type() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("num", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_cast_type",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Cast Int32 to Int64 (a supported cast)
+        table
+            .alter_columns(&[ColumnAlteration::new("num".into()).cast_to(DataType::Int64)])
+            .await
+            .unwrap();
+
+        // Verify type changed
+        let schema = table.schema().await.unwrap();
+        assert_eq!(
+            schema.field_with_name("num").unwrap().data_type(),
+            &DataType::Int64
+        );
+
+        // Query the data and verify the returned type is correct
+        let batches = table
+            .query()
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+        let batch = &batches[0];
+        let values = batch["num"]
+            .as_any()
+            .downcast_ref::<arrow_array::Int64Array>()
+            .unwrap()
+            .values();
+        assert_eq!(values.as_ref(), &[1i64, 2, 3]);
+    }
+
+    #[tokio::test]
+    async fn test_alter_column_invalid_cast_fails() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("num", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_invalid_cast",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Casting Int32 to Float64 is not supported
+        let result = table
+            .alter_columns(&[ColumnAlteration::new("num".into()).cast_to(DataType::Float64)])
+            .await;
+        let err = result.unwrap_err();
+        assert!(
+            err.to_string().contains("cast"),
+            "Expected error message to contain 'cast', got: {}",
+            err
+        );
+    }
+
+    #[tokio::test]
+    async fn test_alter_multiple_columns() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("a", Int32, [1, 2, 3]), ("b", Int32, [4, 5, 6])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_alter_multi",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Alter multiple columns at once
+        table
+            .alter_columns(&[
+                ColumnAlteration::new("a".into()).rename("alpha".into()),
+                ColumnAlteration::new("b".into()).set_nullable(true),
+            ])
+            .await
+            .unwrap();
+
+        let schema = table.schema().await.unwrap();
+        assert!(schema.field_with_name("alpha").is_ok());
+        assert!(schema.field_with_name("a").is_err());
+        assert!(schema.field_with_name("b").unwrap().is_nullable());
+    }
+
+    // Drop Columns Tests
+
+    #[tokio::test]
+    async fn test_drop_single_column() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch =
+            record_batch!(("keep", Int32, [1, 2, 3]), ("remove", Int32, [4, 5, 6])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_drop_single",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        let initial_version = table.version().await.unwrap();
+
+        // Drop a column
+        let result = table.drop_columns(&["remove"]).await.unwrap();
+
+        // Version should increment
+        assert!(result.version > initial_version);
+
+        // Verify column was dropped
+        let schema = table.schema().await.unwrap();
+        assert_eq!(schema.fields().len(), 1);
+        assert!(schema.field_with_name("keep").is_ok());
+        assert!(schema.field_with_name("remove").is_err());
+    }
+
+    #[tokio::test]
+    async fn test_drop_multiple_columns() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(
+            ("a", Int32, [1, 2]),
+            ("b", Int32, [3, 4]),
+            ("c", Int32, [5, 6]),
+            ("d", Int32, [7, 8])
+        )
+        .unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_drop_multi",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Drop multiple columns
+        table.drop_columns(&["b", "d"]).await.unwrap();
+
+        // Verify only a and c remain
+        let schema = table.schema().await.unwrap();
+        assert_eq!(schema.fields().len(), 2);
+        assert!(schema.field_with_name("a").is_ok());
+        assert!(schema.field_with_name("c").is_ok());
+        assert!(schema.field_with_name("b").is_err());
+        assert!(schema.field_with_name("d").is_err());
+    }
+
+    #[tokio::test]
+    async fn test_drop_column_preserves_data() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(
+            ("id", Int32, [1, 2, 3]),
+            ("name", Utf8, ["a", "b", "c"]),
+            ("extra", Int32, [10, 20, 30])
+        )
+        .unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_drop_preserves",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Drop the extra column
+        table.drop_columns(&["extra"]).await.unwrap();
+
+        // Verify remaining data is intact
+        let batches = table
+            .query()
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        let batch = &batches[0];
+        assert_eq!(batch.num_columns(), 2);
+        assert_eq!(batch.num_rows(), 3);
+
+        let ids: Vec<i32> = batch
+            .column(0)
+            .as_any()
+            .downcast_ref::<Int32Array>()
+            .unwrap()
+            .iter()
+            .map(|v| v.unwrap())
+            .collect();
+        assert_eq!(ids, vec![1, 2, 3]);
+
+        let names: Vec<&str> = batch
+            .column(1)
+            .as_any()
+            .downcast_ref::<StringArray>()
+            .unwrap()
+            .iter()
+            .map(|v| v.unwrap())
+            .collect();
+        assert_eq!(names, vec!["a", "b", "c"]);
+    }
+
+    // Error Case Tests
+
+    #[tokio::test]
+    async fn test_drop_nonexistent_column_fails() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("existing", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_drop_nonexistent",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Try to drop a column that doesn't exist
+        let result = table.drop_columns(&["nonexistent"]).await;
+        let err = result.unwrap_err();
+        assert!(
+            err.to_string().contains("nonexistent"),
+            "Expected error message to contain column name 'nonexistent', got: {}",
+            err
+        );
+    }
+
+    #[tokio::test]
+    async fn test_alter_nonexistent_column_fails() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("existing", Int32, [1, 2, 3])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_alter_nonexistent",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        // Try to alter a column that doesn't exist
+        let result = table
+            .alter_columns(&[ColumnAlteration::new("nonexistent".into()).rename("new".into())])
+            .await;
+        let err = result.unwrap_err();
+        assert!(
+            err.to_string().contains("nonexistent"),
+            "Expected error message to contain column name 'nonexistent', got: {}",
+            err
+        );
+    }
+
+    // Version Tracking Tests
+
+    #[tokio::test]
+    async fn test_schema_operations_increment_version() {
+        let conn = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("a", Int32, [1, 2, 3]), ("b", Int32, [4, 5, 6])).unwrap();
+        let schema = batch.schema();
+
+        let table = conn
+            .create_table(
+                "test_version_increment",
+                RecordBatchIterator::new(vec![Ok(batch)], schema),
+            )
+            .execute()
+            .await
+            .unwrap();
+
+        let v1 = table.version().await.unwrap();
+
+        // Add column increments version
+        let add_result = table
+            .add_columns(
+                NewColumnTransform::SqlExpressions(vec![("c".into(), "a + b".into())]),
+                None,
+            )
+            .await
+            .unwrap();
+        assert!(add_result.version > v1);
+        let v2 = table.version().await.unwrap();
+        assert_eq!(add_result.version, v2);
+
+        // Alter column increments version
+        let alter_result = table
+            .alter_columns(&[ColumnAlteration::new("c".into()).rename("sum".into())])
+            .await
+            .unwrap();
+        assert!(alter_result.version > v2);
+        let v3 = table.version().await.unwrap();
+        assert_eq!(alter_result.version, v3);
+
+        // Drop column increments version
+        let drop_result = table.drop_columns(&["b"]).await.unwrap();
+        assert!(drop_result.version > v3);
+        let v4 = table.version().await.unwrap();
+        assert_eq!(drop_result.version, v4);
+    }
+}

rust/lancedb/src/table/update.rs

@@ -0,0 +1,441 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+use std::sync::Arc;
+
+use lance::dataset::UpdateBuilder as LanceUpdateBuilder;
+use serde::{Deserialize, Serialize};
+
+use super::{BaseTable, NativeTable};
+use crate::Error;
+use crate::Result;
+
+/// The result of an update operation
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize, Default)]
+pub struct UpdateResult {
+    #[serde(default)]
+    pub rows_updated: u64,
+    /// The commit version associated with the operation.
+    #[serde(default)]
+    pub version: u64,
+}
+
+/// A builder for configuring a [`crate::table::Table::update`] operation
+#[derive(Debug, Clone)]
+pub struct UpdateBuilder {
+    parent: Arc<dyn BaseTable>,
+    pub(crate) filter: Option<String>,
+    pub(crate) columns: Vec<(String, String)>,
+}
+
+impl UpdateBuilder {
+    pub(crate) fn new(parent: Arc<dyn BaseTable>) -> Self {
+        Self {
+            parent,
+            filter: None,
+            columns: Vec::new(),
+        }
+    }
+
+    /// Limits the update operation to rows matching the given filter
+    ///
+    /// If a row does not match the filter then it will be left unchanged.
+    pub fn only_if(mut self, filter: impl Into<String>) -> Self {
+        self.filter = Some(filter.into());
+        self
+    }
+
+    /// Specifies a column to update
+    ///
+    /// This method may be called multiple times to update multiple columns
+    ///
+    /// The `update_expr` should be an SQL expression explaining how to calculate
+    /// the new value for the column.  The expression will be evaluated against the
+    /// previous row's value.
+    pub fn column(
+        mut self,
+        column_name: impl Into<String>,
+        update_expr: impl Into<String>,
+    ) -> Self {
+        self.columns.push((column_name.into(), update_expr.into()));
+        self
+    }
+
+    /// Executes the update operation.
+    pub async fn execute(self) -> Result<UpdateResult> {
+        if self.columns.is_empty() {
+            Err(Error::InvalidInput {
+                message: "at least one column must be specified in an update operation".to_string(),
+            })
+        } else {
+            self.parent.clone().update(self).await
+        }
+    }
+}
+
+/// Internal implementation of the update logic
+pub(crate) async fn execute_update(
+    table: &NativeTable,
+    update: UpdateBuilder,
+) -> Result<UpdateResult> {
+    // 1. Snapshot the current dataset
+    let dataset = table.dataset.get().await?.clone();
+
+    // 2. Initialize the Lance Core builder
+    let mut builder = LanceUpdateBuilder::new(Arc::new(dataset));
+
+    // 3. Apply the filter (WHERE clause)
+    if let Some(predicate) = update.filter {
+        builder = builder.update_where(&predicate)?;
+    }
+
+    // 4. Apply the columns (SET clause)
+    for (column, value) in update.columns {
+        builder = builder.set(column, &value)?;
+    }
+
+    // 5. Execute the operation (Write new files)
+    let operation = builder.build()?;
+    let res = operation.execute().await?;
+
+    // 6. Update the table's view of the latest version
+    table
+        .dataset
+        .set_latest(res.new_dataset.as_ref().clone())
+        .await;
+
+    Ok(UpdateResult {
+        rows_updated: res.rows_updated,
+        version: res.new_dataset.version().version,
+    })
+}
+
+#[cfg(test)]
+mod tests {
+    use crate::connect;
+    use crate::query::QueryBase;
+    use crate::query::{ExecutableQuery, Select};
+    use arrow_array::{
+        record_batch, Array, BooleanArray, Date32Array, FixedSizeListArray, Float32Array,
+        Float64Array, Int32Array, Int64Array, LargeStringArray, RecordBatch, RecordBatchIterator,
+        RecordBatchReader, StringArray, TimestampMillisecondArray, TimestampNanosecondArray,
+        UInt32Array,
+    };
+    use arrow_data::ArrayDataBuilder;
+    use arrow_schema::{ArrowError, DataType, Field, Schema, TimeUnit};
+    use futures::TryStreamExt;
+    use std::sync::Arc;
+    use std::time::Duration;
+
+    #[tokio::test]
+    async fn test_update_all_types() {
+        let conn = connect("memory://")
+            .read_consistency_interval(Duration::from_secs(0))
+            .execute()
+            .await
+            .unwrap();
+
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("int32", DataType::Int32, false),
+            Field::new("int64", DataType::Int64, false),
+            Field::new("uint32", DataType::UInt32, false),
+            Field::new("string", DataType::Utf8, false),
+            Field::new("large_string", DataType::LargeUtf8, false),
+            Field::new("float32", DataType::Float32, false),
+            Field::new("float64", DataType::Float64, false),
+            Field::new("bool", DataType::Boolean, false),
+            Field::new("date32", DataType::Date32, false),
+            Field::new(
+                "timestamp_ns",
+                DataType::Timestamp(TimeUnit::Nanosecond, None),
+                false,
+            ),
+            Field::new(
+                "timestamp_ms",
+                DataType::Timestamp(TimeUnit::Millisecond, None),
+                false,
+            ),
+            Field::new(
+                "vec_f32",
+                DataType::FixedSizeList(Arc::new(Field::new("item", DataType::Float32, true)), 2),
+                false,
+            ),
+            Field::new(
+                "vec_f64",
+                DataType::FixedSizeList(Arc::new(Field::new("item", DataType::Float64, true)), 2),
+                false,
+            ),
+        ]));
+
+        let record_batch_iter = RecordBatchIterator::new(
+            vec![RecordBatch::try_new(
+                schema.clone(),
+                vec![
+                    Arc::new(Int32Array::from_iter_values(0..10)),
+                    Arc::new(Int64Array::from_iter_values(0..10)),
+                    Arc::new(UInt32Array::from_iter_values(0..10)),
+                    Arc::new(StringArray::from_iter_values(vec![
+                        "a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
+                    ])),
+                    Arc::new(LargeStringArray::from_iter_values(vec![
+                        "a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
+                    ])),
+                    Arc::new(Float32Array::from_iter_values((0..10).map(|i| i as f32))),
+                    Arc::new(Float64Array::from_iter_values((0..10).map(|i| i as f64))),
+                    Arc::new(Into::<BooleanArray>::into(vec![
+                        true, false, true, false, true, false, true, false, true, false,
+                    ])),
+                    Arc::new(Date32Array::from_iter_values(0..10)),
+                    Arc::new(TimestampNanosecondArray::from_iter_values(0..10)),
+                    Arc::new(TimestampMillisecondArray::from_iter_values(0..10)),
+                    Arc::new(
+                        create_fixed_size_list(
+                            Float32Array::from_iter_values((0..20).map(|i| i as f32)),
+                            2,
+                        )
+                        .unwrap(),
+                    ),
+                    Arc::new(
+                        create_fixed_size_list(
+                            Float64Array::from_iter_values((0..20).map(|i| i as f64)),
+                            2,
+                        )
+                        .unwrap(),
+                    ),
+                ],
+            )
+            .unwrap()]
+            .into_iter()
+            .map(Ok),
+            schema.clone(),
+        );
+
+        let table = conn
+            .create_table("my_table", record_batch_iter)
+            .execute()
+            .await
+            .unwrap();
+
+        // check it can do update for each type
+        let updates: Vec<(&str, &str)> = vec![
+            ("string", "'foo'"),
+            ("large_string", "'large_foo'"),
+            ("int32", "1"),
+            ("int64", "1"),
+            ("uint32", "1"),
+            ("float32", "1.0"),
+            ("float64", "1.0"),
+            ("bool", "true"),
+            ("date32", "1"),
+            ("timestamp_ns", "1"),
+            ("timestamp_ms", "1"),
+            ("vec_f32", "[1.0, 1.0]"),
+            ("vec_f64", "[1.0, 1.0]"),
+        ];
+
+        let mut update_op = table.update();
+        for (column, value) in updates {
+            update_op = update_op.column(column, value);
+        }
+        update_op.execute().await.unwrap();
+
+        let mut batches = table
+            .query()
+            .select(Select::columns(&[
+                "string",
+                "large_string",
+                "int32",
+                "int64",
+                "uint32",
+                "float32",
+                "float64",
+                "bool",
+                "date32",
+                "timestamp_ns",
+                "timestamp_ms",
+                "vec_f32",
+                "vec_f64",
+            ]))
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+        let batch = batches.pop().unwrap();
+
+        macro_rules! assert_column {
+            ($column:expr, $array_type:ty, $expected:expr) => {
+                let array = $column
+                    .as_any()
+                    .downcast_ref::<$array_type>()
+                    .unwrap()
+                    .iter()
+                    .collect::<Vec<_>>();
+                for v in array {
+                    assert_eq!(v, Some($expected));
+                }
+            };
+        }
+
+        assert_column!(batch.column(0), StringArray, "foo");
+        assert_column!(batch.column(1), LargeStringArray, "large_foo");
+        assert_column!(batch.column(2), Int32Array, 1);
+        assert_column!(batch.column(3), Int64Array, 1);
+        assert_column!(batch.column(4), UInt32Array, 1);
+        assert_column!(batch.column(5), Float32Array, 1.0);
+        assert_column!(batch.column(6), Float64Array, 1.0);
+        assert_column!(batch.column(7), BooleanArray, true);
+        assert_column!(batch.column(8), Date32Array, 1);
+        assert_column!(batch.column(9), TimestampNanosecondArray, 1);
+        assert_column!(batch.column(10), TimestampMillisecondArray, 1);
+
+        let array = batch
+            .column(11)
+            .as_any()
+            .downcast_ref::<FixedSizeListArray>()
+            .unwrap()
+            .iter()
+            .collect::<Vec<_>>();
+        for v in array {
+            let v = v.unwrap();
+            let f32array = v.as_any().downcast_ref::<Float32Array>().unwrap();
+            for v in f32array {
+                assert_eq!(v, Some(1.0));
+            }
+        }
+
+        let array = batch
+            .column(12)
+            .as_any()
+            .downcast_ref::<FixedSizeListArray>()
+            .unwrap()
+            .iter()
+            .collect::<Vec<_>>();
+        for v in array {
+            let v = v.unwrap();
+            let f64array = v.as_any().downcast_ref::<Float64Array>().unwrap();
+            for v in f64array {
+                assert_eq!(v, Some(1.0));
+            }
+        }
+    }
+    ///Two helper functions
+    fn create_fixed_size_list<T: Array>(
+        values: T,
+        list_size: i32,
+    ) -> Result<FixedSizeListArray, ArrowError> {
+        let list_type = DataType::FixedSizeList(
+            Arc::new(Field::new("item", values.data_type().clone(), true)),
+            list_size,
+        );
+        let data = ArrayDataBuilder::new(list_type)
+            .len(values.len() / list_size as usize)
+            .add_child_data(values.into_data())
+            .build()
+            .unwrap();
+
+        Ok(FixedSizeListArray::from(data))
+    }
+
+    fn make_test_batches() -> impl RecordBatchReader + Send + Sync + 'static {
+        let schema = Arc::new(Schema::new(vec![Field::new("i", DataType::Int32, false)]));
+        RecordBatchIterator::new(
+            vec![RecordBatch::try_new(
+                schema.clone(),
+                vec![Arc::new(Int32Array::from_iter_values(0..10))],
+            )],
+            schema,
+        )
+    }
+
+    #[tokio::test]
+    async fn test_update_with_predicate() {
+        let conn = connect("memory://")
+            .read_consistency_interval(Duration::from_secs(0))
+            .execute()
+            .await
+            .unwrap();
+
+        let batch = record_batch!(
+            ("id", Int32, [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]),
+            (
+                "name",
+                Utf8,
+                ["a", "b", "c", "d", "e", "f", "g", "h", "i", "j"]
+            )
+        )
+        .unwrap();
+
+        let schema = batch.schema();
+        // need the iterator for create table
+        let record_batch_iter = RecordBatchIterator::new(vec![Ok(batch)], schema);
+
+        let table = conn
+            .create_table("my_table", record_batch_iter)
+            .execute()
+            .await
+            .unwrap();
+
+        table
+            .update()
+            .only_if("id > 5")
+            .column("name", "'foo'")
+            .execute()
+            .await
+            .unwrap();
+
+        let mut batches = table
+            .query()
+            .select(Select::columns(&["id", "name"]))
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        while let Some(batch) = batches.pop() {
+            let ids = batch
+                .column(0)
+                .as_any()
+                .downcast_ref::<Int32Array>()
+                .unwrap()
+                .iter()
+                .collect::<Vec<_>>();
+            let names = batch
+                .column(1)
+                .as_any()
+                .downcast_ref::<StringArray>()
+                .unwrap()
+                .iter()
+                .collect::<Vec<_>>();
+            for (i, name) in names.iter().enumerate() {
+                let id = ids[i].unwrap();
+                let name = name.unwrap();
+                if id > 5 {
+                    assert_eq!(name, "foo");
+                } else {
+                    assert_eq!(name, &format!("{}", (b'a' + id as u8) as char));
+                }
+            }
+        }
+    }
+
+    #[tokio::test]
+    async fn test_update_via_expr() {
+        let conn = connect("memory://")
+            .read_consistency_interval(Duration::from_secs(0))
+            .execute()
+            .await
+            .unwrap();
+        let tbl = conn
+            .create_table("my_table", make_test_batches())
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(1, tbl.count_rows(Some("i == 0".to_string())).await.unwrap());
+        tbl.update().column("i", "i+1").execute().await.unwrap();
+        assert_eq!(0, tbl.count_rows(Some("i == 0".to_string())).await.unwrap());
+    }
+}

rust/lancedb/tests/embedding_registry_test.rs

@@ -4,7 +4,6 @@
 use std::{
     borrow::Cow,
     collections::{HashMap, HashSet},
-    iter::repeat,
     sync::Arc,
 };
 
@@ -268,9 +267,10 @@ fn create_some_records() -> Result<impl IntoArrow> {
             schema.clone(),
             vec![
                 Arc::new(Int32Array::from_iter_values(0..TOTAL as i32)),
-                Arc::new(StringArray::from_iter(
-                    repeat(Some("hello world".to_string())).take(TOTAL),
-                )),
+                Arc::new(StringArray::from_iter(std::iter::repeat_n(
+                    Some("hello world".to_string()),
+                    TOTAL,
+                ))),
             ],
         )
         .unwrap()]