chore(deps): bump snafu from 0.8.9 to 0.9.0

Bumps [snafu](https://github.com/shepmaster/snafu) from 0.8.9 to 0.9.0.
- [Changelog](https://github.com/shepmaster/snafu/blob/main/CHANGELOG.md)
- [Commits](https://github.com/shepmaster/snafu/compare/0.8.9...0.9.0)

---
updated-dependencies:
- dependency-name: snafu
  dependency-version: 0.9.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.28.0-beta.11"
+current_version = "0.29.1-beta.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/dependabot.yml

@@ -11,6 +11,11 @@ updates:
     schedule:
       interval: weekly
     open-pull-requests-limit: 10
+    # Only update Cargo.lock, never widen/raise the version requirements in
+    # Cargo.toml. The goal is keeping the lockfile (and the binaries we ship)
+    # current on security fixes, not forcing our library's consumers onto
+    # newer minimum versions.
+    versioning-strategy: lockfile-only
     groups:
       rust-minor-patch:
         update-types:

.github/workflows/codex-fix-ci.yml

@@ -45,7 +45,9 @@ jobs:
       - name: Set up Node.js
         uses: actions/setup-node@v4
         with:
-          node-version: 20
+          # pnpm 11 (used by the nodejs install step below) requires
+          # Node >= 22.13; use 24 since 22 hits EOL in October.
+          node-version: 24
 
       - name: Install Codex CLI
         run: npm install -g @openai/codex
@@ -79,10 +81,14 @@ jobs:
           java-version: '11'
           cache: maven
 
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+        with:
+          version: 11.1.1
       - name: Install Node.js dependencies for TypeScript bindings
         run: |
           cd nodejs
-          npm ci
+          pnpm install --frozen-lockfile
 
       - name: Configure git user
         run: |
@@ -137,7 +143,7 @@ jobs:
                - For Rust test failures: Run the specific test with "cargo test -p <crate> <test_name>"
                - For Python test failures: Build with "cd python && maturin develop" then run "pytest <specific_test_file>::<test_name>"
                - For Java test failures: Run "cd java && mvn test -Dtest=<TestClass>#<testMethod>"
-               - For TypeScript test failures: Run "cd nodejs && npm run build && npm test -- --testNamePattern='<test_name>'"
+               - For TypeScript test failures: Run "cd nodejs && pnpm build && pnpm test -- --testNamePattern='<test_name>'"
                - Do NOT run the full test suite - only run the tests that were failing
 
           7. If the additional guidelines are provided, follow them as well.

.github/workflows/nodejs.yml

@@ -42,11 +42,17 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
+    - uses: pnpm/action-setup@v4
+      with:
+        version: 11.1.1
     - uses: actions/setup-node@v4
       with:
-        node-version: 20
-        cache: 'npm'
-        cache-dependency-path: nodejs/package-lock.json
+        # pnpm 11 requires Node >= 22.13; use 24 since 22 hits EOL
+        # in October. The library itself still supports Node >= 18
+        # (see test matrix below).
+        node-version: 24
+        cache: 'pnpm'
+        cache-dependency-path: nodejs/pnpm-lock.yaml
     - uses: actions-rust-lang/setup-rust-toolchain@v1
       with:
         components: rustfmt, clippy
@@ -61,11 +67,13 @@ jobs:
       run: cargo clippy --profile ci --all --all-features -- -D warnings
     - name: Lint Typescript
       run: |
-        npm ci
-        npm run lint-ci
+        pnpm install --frozen-lockfile
+        pnpm lint-ci
     - name: Lint examples
       working-directory: nodejs/examples
-      run: npm ci && npm run lint-ci
+      # The `@lancedb/lancedb` dep points at file:../dist; pnpm errors if
+      # that dir is missing, so create an empty one for lint-only runs.
+      run: mkdir -p ../dist && pnpm install --frozen-lockfile && pnpm lint-ci
   linux:
     name: Linux (NodeJS ${{ matrix.node-version }})
     timeout-minutes: 30
@@ -82,14 +90,18 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
-    - uses: actions/setup-node@v4
-      name: Setup Node.js 20 for build
+    - uses: pnpm/action-setup@v4
       with:
-        # @napi-rs/cli v3 requires Node >= 20.12 (via @inquirer/prompts@8).
-        # Build always on Node 20; tests run on the matrix version below.
-        node-version: 20
-        cache: 'npm'
-        cache-dependency-path: nodejs/package-lock.json
+        version: 11.1.1
+    - uses: actions/setup-node@v4
+      name: Setup Node.js 24 for build
+      with:
+        # pnpm 11 requires Node >= 22.13; use 24 since 22 hits EOL
+        # in October. Build/install runs on Node 24; tests run on the
+        # matrix version below using direct jest invocation.
+        node-version: 24
+        cache: 'pnpm'
+        cache-dependency-path: nodejs/pnpm-lock.yaml
     - uses: Swatinem/rust-cache@v2
     - name: Install dependencies
       run: |
@@ -97,48 +109,58 @@ jobs:
         sudo apt install -y protobuf-compiler libssl-dev
     - name: Build
       run: |
-        npm ci --include=optional
-        npm run build:debug -- --profile ci
+        pnpm install --frozen-lockfile
+        # No `--` separator: pnpm forwards it literally, which would
+        # make napi-rs treat `--profile ci` as a cargo passthrough arg.
+        pnpm build:debug --profile ci
+        pnpm tsc
+    - name: Setup examples
+      working-directory: nodejs/examples
+      run: pnpm install --frozen-lockfile
+    - name: Check docs
+      run: |
+        # We run this as part of the job because the binary needs to be built
+        # first to export the types of the native code.
+        set -e
+        # `pnpm docs` would invoke pnpm's built-in `docs` command, not
+        # the script — use `pnpm run docs`.
+        pnpm run docs
+        if ! git diff --exit-code -- ../ ':(exclude)Cargo.lock'; then
+          echo "Docs need to be updated"
+          echo "Run 'pnpm run docs', fix any warnings, and commit the changes."
+          exit 1
+        fi
     - uses: actions/setup-node@v4
       name: Setup Node.js ${{ matrix.node-version }} for test
       with:
         node-version: ${{ matrix.node-version }}
-    - name: Compile TypeScript
-      run: npm run tsc
     - name: Setup localstack
       working-directory: .
       run: docker compose up --detach --wait
     - name: Test
       env:
         S3_TEST: "1"
-      run: npm run test
-    - name: Setup examples
-      working-directory: nodejs/examples
-      run: npm ci
+        # Newer @smithy/core uses dynamic ESM imports.
+        NODE_OPTIONS: "--experimental-vm-modules"
+      # Invoke jest directly because pnpm 11 itself requires Node 22+
+      # while the matrix tests on older Node versions.
+      run: npx jest --verbose
     - name: Test examples
       working-directory: ./
       env:
         OPENAI_API_KEY: test
         OPENAI_BASE_URL: http://0.0.0.0:8000
+        NODE_OPTIONS: "--experimental-vm-modules"
       run: |
         python ci/mock_openai.py &
         cd nodejs/examples
-        npm test
-    - name: Check docs
-      run: |
-        # We run this as part of the job because the binary needs to be built
-        # first to export the types of the native code.
-        set -e
-        npm ci
-        npm run docs
-        if ! git diff --exit-code -- ../ ':(exclude)Cargo.lock'; then
-          echo "Docs need to be updated"
-          echo "Run 'npm run docs', fix any warnings, and commit the changes."
-          exit 1
-        fi
+        npx jest --testEnvironment jest-environment-node-single-context --verbose
   macos:
     timeout-minutes: 30
-    runs-on: "macos-14"
+    # macos-15 ships a newer linker; the older macos-14 linker fails to insert
+    # branch islands when the debug cdylib's __text section exceeds the 128 MB
+    # AArch64 B/BL branch range.
+    runs-on: "macos-15"
     defaults:
       run:
         shell: bash
@@ -148,20 +170,28 @@ jobs:
       with:
         fetch-depth: 0
         lfs: true
+    - uses: pnpm/action-setup@v4
+      with:
+        version: 11.1.1
     - uses: actions/setup-node@v4
       with:
-        node-version: 20
-        cache: 'npm'
-        cache-dependency-path: nodejs/package-lock.json
+        # pnpm 11 requires Node >= 22.13; use 24 since 22 hits EOL
+        # in October.
+        node-version: 24
+        cache: 'pnpm'
+        cache-dependency-path: nodejs/pnpm-lock.yaml
+    - uses: dtolnay/rust-toolchain@stable
     - uses: Swatinem/rust-cache@v2
     - name: Install dependencies
       run: |
         brew install protobuf
     - name: Build
       run: |
-        npm ci --include=optional
-        npm run build:debug -- --profile ci
-        npm run tsc
+        pnpm install --frozen-lockfile
+        # No `--` separator: pnpm forwards it literally, which would
+        # make napi-rs treat `--profile ci` as a cargo passthrough arg.
+        pnpm build:debug --profile ci
+        pnpm tsc
     - name: Test
       run: |
-        npm run test
+        pnpm test

.github/workflows/npm-publish.yml

@@ -171,13 +171,18 @@ jobs:
         working-directory: nodejs
     steps:
       - uses: actions/checkout@v4
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+        with:
+          version: 11.1.1
       - name: Setup node
         uses: actions/setup-node@v4
-        if: ${{ !matrix.settings.docker }}
         with:
-          node-version: 20
-          cache: npm
-          cache-dependency-path: nodejs/package-lock.json
+          # pnpm 11 requires Node >= 22.13; use 24 since 22 hits EOL
+          # in October.
+          node-version: 24
+          cache: pnpm
+          cache-dependency-path: nodejs/pnpm-lock.yaml
       - name: Install
         uses: dtolnay/rust-toolchain@stable
         if: ${{ !matrix.settings.docker }}
@@ -195,7 +200,7 @@ jobs:
             target/
           key: nodejs-${{ matrix.settings.target }}-cargo-${{ matrix.settings.host }}
       - name: Install dependencies
-        run: npm ci
+        run: pnpm install --frozen-lockfile
       - name: Install Zig
         uses: mlugg/setup-zig@v2
         if: ${{ contains(matrix.settings.target, 'musl') }}
@@ -248,7 +253,7 @@ jobs:
       # one to do the upload.
       - name: Make generic artifacts
         if: ${{ matrix.settings.target == 'aarch64-apple-darwin' }}
-        run: npm run tsc
+        run: pnpm tsc
       - name: Upload Generic Artifacts
         if: ${{ matrix.settings.target == 'aarch64-apple-darwin' }}
         uses: actions/upload-artifact@v4
@@ -283,14 +288,24 @@ jobs:
         working-directory: nodejs
     steps:
       - uses: actions/checkout@v4
-      - name: Setup node
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+        with:
+          version: 11.1.1
+      - name: Setup Node.js 24 for install
+        uses: actions/setup-node@v4
+        with:
+          # pnpm 11 requires Node >= 22.13; use 24 since 22 hits EOL
+          # in October.
+          node-version: 24
+          cache: pnpm
+          cache-dependency-path: nodejs/pnpm-lock.yaml
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile
+      - name: Setup Node.js ${{ matrix.node }} for test
         uses: actions/setup-node@v4
         with:
           node-version: ${{ matrix.node }}
-          cache: npm
-          cache-dependency-path: nodejs/package-lock.json
-      - name: Install dependencies
-        run: npm ci
       - name: Download artifacts
         uses: actions/download-artifact@v4
         with:
@@ -311,7 +326,9 @@ jobs:
       - name: Move built files
         run: cp dist/native.d.ts dist/native.js dist/*.node lancedb/
       - name: Test bindings
-        run: npm test
+        # Invoke jest directly because pnpm 11 itself requires Node 22+
+        # while the matrix tests on older Node versions.
+        run: npx jest --verbose
   publish:
     name: Publish
     runs-on: ubuntu-latest
@@ -323,15 +340,19 @@ jobs:
       - test-lancedb
     steps:
       - uses: actions/checkout@v4
+      - name: Setup pnpm
+        uses: pnpm/action-setup@v4
+        with:
+          version: 11.1.1
       - name: Setup node
         uses: actions/setup-node@v4
         with:
           node-version: 24
-          cache: npm
-          cache-dependency-path: nodejs/package-lock.json
+          cache: pnpm
+          cache-dependency-path: nodejs/pnpm-lock.yaml
           registry-url: "https://registry.npmjs.org"
       - name: Install dependencies
-        run: npm ci
+        run: pnpm install --frozen-lockfile
       - uses: actions/download-artifact@v4
         with:
           name: nodejs-dist
@@ -351,7 +372,7 @@ jobs:
       - name: Display structure of downloaded files
         run: find dist && find nodejs-artifacts
       - name: Move artifacts
-        run: npx napi artifacts -d nodejs-artifacts
+        run: pnpm exec napi artifacts -d nodejs-artifacts
       - name: List packages
         run: find npm
       - name: Publish

.github/workflows/python.yml

@@ -205,7 +205,7 @@ jobs:
       - name: Delete wheels
         run: rm -rf target/wheels
   pydantic1x:
-    timeout-minutes: 30
+    timeout-minutes: 60
     runs-on: "ubuntu-24.04"
     defaults:
       run:

.github/workflows/rust.yml

@@ -233,6 +233,26 @@ jobs:
           cargo update -p aws-sdk-sso --precise 1.62.0
           cargo update -p aws-sdk-ssooidc --precise 1.63.0
           cargo update -p aws-sdk-sts --precise 1.63.0
+          # aws-runtime/sigv4/credential-types/types and the aws-smithy-*
+          # crates bumped their MSRV to 1.91.1 in late 2026; pin to the last
+          # 1.91.0-compatible versions. The order matters — each downgrade
+          # only succeeds once everything that still pins it at a higher
+          # version has itself been downgraded.
+          cargo update -p aws-runtime --precise 1.5.12
+          cargo update -p aws-types --precise 1.3.9
+          cargo update -p aws-sigv4 --precise 1.3.5
+          cargo update -p aws-credential-types --precise 1.2.8
+          cargo update -p aws-smithy-checksums --precise 0.63.9
+          cargo update -p aws-smithy-runtime --precise 1.9.3
+          cargo update -p aws-smithy-http --precise 0.62.4
+          cargo update -p aws-smithy-eventstream --precise 0.60.12
+          cargo update -p aws-smithy-http-client --precise 1.1.3
+          cargo update -p aws-smithy-observability --precise 0.1.4
+          cargo update -p aws-smithy-query --precise 0.60.8
+          cargo update -p aws-smithy-runtime-api --precise 1.9.1
+          cargo update -p aws-smithy-async --precise 1.2.6
+          cargo update -p aws-smithy-types --precise 1.3.5
+          cargo update -p aws-smithy-xml --precise 0.60.11
           cargo update -p home --precise 0.5.9
       - name: cargo +${{ matrix.msrv }} check
         env:

AGENTS.md

@@ -17,9 +17,30 @@ Common commands:
 * Run tests: `cargo test --quiet --features remote --tests`
 * Run specific test: `cargo test --quiet --features remote -p <package_name> --test <test_name>`
 * Lint: `cargo clippy --quiet --features remote --tests --examples`
-* Format: `cargo fmt --all`
+* Format Rust: `cargo fmt --all`
+* Format Python: `ruff format .`
+* Lint Python: `ruff check .`
+* Bootstrap Python dev env: `cd python && uv run --extra tests --extra dev maturin develop --extras tests,dev`
+* Run Python tests: `cd python && uv run --extra tests pytest python/tests -vv --durations=10 -m "not slow and not s3_test"`
+* Run specific Python test: `cd python && uv run --extra tests pytest python/tests/<test_file>.py::<test_name> -q`
 
-Before committing changes, run formatting.
+For Python validation, prefer the uv-managed environment declared by `python/uv.lock`.
+Do not treat system `python`, global `pytest`, or missing editable-install errors as
+final blockers; bootstrap or enter the uv environment instead. If `lancedb._lancedb`
+is missing or stale, or if Rust/PyO3 binding code changed, rebuild the Python
+extension with the bootstrap command above before running tests.
+
+Before committing changes, run formatting for every language you touched. At minimum:
+
+* Rust changes: run `cargo fmt --all`.
+* Python changes: run `ruff format .` and `ruff check .` from the repository root,
+  and run targeted tests through `cd python && uv run ...`.
+* TypeScript changes: run the relevant `npm`/`pnpm` lint, format, build, and docs commands in `nodejs`.
+
+Before creating a PR, make sure the PR title follows Conventional Commits, such as
+`fix: support nested field paths in native index creation` or
+`feat(python): add dataset multiprocessing support`. The semantic-release check uses the
+PR title and body as the merge commit message, so a non-conventional PR title will fail CI.
 
 ## Coding tips
 

Cargo.toml

@@ -13,20 +13,20 @@ categories = ["database-implementations"]
 rust-version = "1.91.0"
 
 [workspace.dependencies]
-lance = { "version" = "=7.0.0-beta.7", default-features = false, "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=7.0.0-beta.7", default-features = false, "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=7.0.0-beta.7", default-features = false, "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=7.0.0-beta.7", "tag" = "v7.0.0-beta.7", "git" = "https://github.com/lance-format/lance.git" }
+lance = { "version" = "=7.0.0-beta.13", default-features = false, "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-core = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-datagen = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-file = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-io = { "version" = "=7.0.0-beta.13", default-features = false, "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-index = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-linalg = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-namespace-impls = { "version" = "=7.0.0-beta.13", default-features = false, "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-table = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-testing = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-datafusion = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-encoding = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
+lance-arrow = { "version" = "=7.0.0-beta.13", "tag" = "v7.0.0-beta.13", "git" = "https://github.com/lance-format/lance.git" }
 ahash = "0.8"
 # Note that this one does not include pyarrow
 arrow = { version = "58.0.0", optional = false }

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>0.28.0-beta.11</version>
+    <version>0.29.1-beta.0</version>
 </dependency>
 ```
 

docs/src/js/_media/CONTRIBUTING.md

@@ -12,20 +12,22 @@ Typescript.
 * `src/`: Rust bindings source code
 * `lancedb/`: Typescript package source code
 * `__test__/`: Unit tests
-* `examples/`: An npm package with the examples shown in the documentation
+* `examples/`: A pnpm package with the examples shown in the documentation
 
 ## Development environment
 
 To set up your development environment, you will need to install the following:
 
-1. Node.js 14 or later
-2. Rust's package manager, Cargo. Use [rustup](https://rustup.rs/) to install.
-3. [protoc](https://grpc.io/docs/protoc-installation/) (Protocol Buffers compiler)
+1. Node.js 22 or later (required by pnpm 11)
+2. [pnpm](https://pnpm.io/installation) 11 or later (or run via `corepack enable`,
+   which uses the `packageManager` field in `package.json`)
+3. Rust's package manager, Cargo. Use [rustup](https://rustup.rs/) to install.
+4. [protoc](https://grpc.io/docs/protoc-installation/) (Protocol Buffers compiler)
 
 Initial setup:
 
 ```shell
-npm install
+pnpm install
 ```
 
 ### Commit Hooks
@@ -39,38 +41,38 @@ pre-commit install
 
 ## Development
 
-Most common development commands can be run using the npm scripts.
+Most common development commands can be run using the pnpm scripts.
 
 Build the package
 
 ```shell
-npm install
-npm run build
+pnpm install
+pnpm build
 ```
 
 Lint:
 
 ```shell
-npm run lint
+pnpm lint
 ```
 
 Format and fix lints:
 
 ```shell
-npm run lint-fix
+pnpm lint-fix
 ```
 
 Run tests:
 
 ```shell
-npm test
+pnpm test
 ```
 
 To run a single test:
 
 ```shell
 # Single file: table.test.ts
-npm test -- table.test.ts
+pnpm test -- table.test.ts
 # Single test: 'merge insert' in table.test.ts
-npm test -- table.test.ts --testNamePattern=merge\ insert
+pnpm test -- table.test.ts --testNamePattern=merge\ insert
 ```

docs/src/js/classes/Connection.md

@@ -148,6 +148,33 @@ Creates a new empty Table
 
 ***
 
+### createNamespace()
+
+```ts
+abstract createNamespace(namespacePath, options?): Promise<CreateNamespaceResponse>
+```
+
+Create a new namespace at the given path.
+
+#### Parameters
+
+* **namespacePath**: `string`[]
+    The namespace path to create.
+
+* **options?**: `Partial`&lt;[`CreateNamespaceOptions`](../interfaces/CreateNamespaceOptions.md)&gt;
+    Creation `mode`
+    ("create" | "exist_ok" | "overwrite") and optional `properties`
+    to attach to the namespace.
+
+#### Returns
+
+`Promise`&lt;[`CreateNamespaceResponse`](../interfaces/CreateNamespaceResponse.md)&gt;
+
+The properties of the
+  created namespace and an optional transaction id.
+
+***
+
 ### createTable()
 
 #### createTable(options, namespacePath)
@@ -230,6 +257,29 @@ Creates a new Table and initialize it with new data.
 
 ***
 
+### describeNamespace()
+
+```ts
+abstract describeNamespace(namespacePath): Promise<DescribeNamespaceResponse>
+```
+
+Describe a namespace, returning its properties.
+
+#### Parameters
+
+* **namespacePath**: `string`[]
+    The namespace path to describe, in
+    parent → child order, e.g. `["analytics", "sales"]`.
+
+#### Returns
+
+`Promise`&lt;[`DescribeNamespaceResponse`](../interfaces/DescribeNamespaceResponse.md)&gt;
+
+The namespace's properties
+  (may be undefined if the namespace has none).
+
+***
+
 ### display()
 
 ```ts
@@ -263,6 +313,36 @@ Drop all tables in the database.
 
 ***
 
+### dropNamespace()
+
+```ts
+abstract dropNamespace(namespacePath, options?): Promise<DropNamespaceResponse>
+```
+
+Drop a namespace.
+
+Use `behavior: "cascade"` to also drop everything contained in the
+namespace (sub-namespaces and tables). The default `"restrict"`
+behavior refuses to drop a non-empty namespace.
+
+#### Parameters
+
+* **namespacePath**: `string`[]
+    The namespace path to drop.
+
+* **options?**: `Partial`&lt;[`DropNamespaceOptions`](../interfaces/DropNamespaceOptions.md)&gt;
+    `mode` ("skip" | "fail"
+    for missing-namespace handling) and `behavior` ("restrict" | "cascade").
+
+#### Returns
+
+`Promise`&lt;[`DropNamespaceResponse`](../interfaces/DropNamespaceResponse.md)&gt;
+
+Any properties returned by
+  the server and an optional transaction id.
+
+***
+
 ### dropTable()
 
 ```ts
@@ -299,6 +379,36 @@ Return true if the connection has not been closed
 
 ***
 
+### listNamespaces()
+
+```ts
+abstract listNamespaces(namespacePath?, options?): Promise<ListNamespacesResponse>
+```
+
+List the immediate child namespaces under the given parent.
+
+Results may be paginated. To retrieve subsequent pages, pass the
+`pageToken` returned by a previous call.
+
+#### Parameters
+
+* **namespacePath?**: `string`[]
+    The parent namespace path. Defaults
+    to the root namespace if omitted.
+
+* **options?**: `Partial`&lt;[`ListNamespacesOptions`](../interfaces/ListNamespacesOptions.md)&gt;
+    Pagination options
+    (`pageToken`, `limit`).
+
+#### Returns
+
+`Promise`&lt;[`ListNamespacesResponse`](../interfaces/ListNamespacesResponse.md)&gt;
+
+Child namespace names and
+  an optional token for fetching the next page.
+
+***
+
 ### openTable()
 
 ```ts
@@ -327,6 +437,39 @@ Open a table in the database.
 
 ***
 
+### renameTable()
+
+```ts
+abstract renameTable(
+   currentName,
+   newName,
+   options?): Promise<void>
+```
+
+Rename a table.
+
+Currently only supported by LanceDB Cloud. Local OSS connections and
+namespace-backed connections (via [connectNamespace](../functions/connectNamespace.md)) reject with
+a "not supported" error.
+
+#### Parameters
+
+* **currentName**: `string`
+    The current name of the table.
+
+* **newName**: `string`
+    The new name for the table.
+
+* **options?**: [`RenameTableOptions`](../interfaces/RenameTableOptions.md)
+    Optional namespace paths. When
+    `newNamespacePath` is omitted the table stays in `namespacePath`.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
 ### tableNames()
 
 #### tableNames(options)

docs/src/js/classes/Query.md

@@ -343,6 +343,30 @@ This is useful for pagination.
 
 ***
 
+### orderBy()
+
+```ts
+orderBy(ordering): this
+```
+
+Sort the results by the specified column(s).
+
+#### Parameters
+
+* **ordering**: [`ColumnOrdering`](../interfaces/ColumnOrdering.md) \| [`ColumnOrdering`](../interfaces/ColumnOrdering.md)[]
+
+#### Returns
+
+`this`
+
+This query builder.
+
+#### Inherited from
+
+`StandardQueryBase.orderBy`
+
+***
+
 ### outputSchema()
 
 ```ts

docs/src/js/classes/Scannable.md

@@ -0,0 +1,173 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / Scannable
+
+# Class: Scannable
+
+A data source that can be scanned as a stream of Arrow `RecordBatch`es.
+
+`Scannable` wraps the schema + optional row count + rescannable flag and
+a callback that yields batches one at a time. It is passed to consumers
+(e.g. `Table.add`, `createTable`, `mergeInsert` — follow-up work) that
+need to pull data without materializing the full dataset in JS memory.
+
+Batches cross the JS↔Rust boundary as Arrow IPC Stream messages; a fresh
+writer serializes each batch, and the Rust side decodes it with
+`arrow_ipc::reader::StreamReader`. One batch is in flight at a time.
+
+## Properties
+
+### numRows
+
+```ts
+readonly numRows: null | number;
+```
+
+***
+
+### rescannable
+
+```ts
+readonly rescannable: boolean;
+```
+
+***
+
+### schema
+
+```ts
+readonly schema: Schema<any>;
+```
+
+## Methods
+
+### fromFactory()
+
+```ts
+static fromFactory(
+   schema,
+   factory,
+   opts): Promise<Scannable>
+```
+
+Build a Scannable from an explicit schema and a factory that returns a
+fresh batch iterator on each call.
+
+The factory is invoked once per scan. Each iterator yields
+`RecordBatch`es matching the declared schema. Use this when you need
+direct control over the pull loop — for example, to wrap a streaming
+source whose batches are produced lazily.
+
+#### Parameters
+
+* **schema**: `Schema`&lt;`any`&gt;
+    The Arrow schema of the produced batches.
+
+* **factory**
+    Called at the start of each scan to produce a batch
+    iterator. Must be idempotent when `rescannable` is true.
+
+* **opts**: [`ScannableOptions`](../interfaces/ScannableOptions.md) = `{}`
+    Optional hints. `rescannable` defaults to `true`; set to
+    `false` if calling `factory()` twice would not reproduce the same data.
+
+#### Returns
+
+`Promise`&lt;[`Scannable`](Scannable.md)&gt;
+
+***
+
+### fromIterable()
+
+```ts
+static fromIterable(
+   schema,
+   iter,
+   opts): Promise<Scannable>
+```
+
+Build a Scannable from an iterable of `RecordBatch`es. `rescannable`
+defaults to `false`. Pass an explicit schema so the consumer can
+validate before any batch is pulled.
+
+`opts.rescannable: true` is honest for replayable iterables (Arrays,
+Sets, or custom iterables whose `[Symbol.iterator]()` returns a fresh
+iterator each call). It is rejected for one-shot iterables (generators,
+async generators, or already-an-iterator inputs) because their
+`[Symbol.iterator]()` returns the same exhausted object on the second
+scan. For replayable sources outside this shape, use
+`fromFactory(schema, () => createIter(), { rescannable: true })`.
+
+Note: when `opts.rescannable` is `true`, the constructor calls
+`[Symbol.iterator]()` once on the input to perform the structural check.
+
+#### Parameters
+
+* **schema**: `Schema`&lt;`any`&gt;
+
+* **iter**: `Iterable`&lt;`RecordBatch`&lt;`any`&gt;&gt; \| `AsyncIterable`&lt;`RecordBatch`&lt;`any`&gt;&gt;
+
+* **opts**: [`ScannableOptions`](../interfaces/ScannableOptions.md) = `{}`
+
+#### Returns
+
+`Promise`&lt;[`Scannable`](Scannable.md)&gt;
+
+***
+
+### fromRecordBatchReader()
+
+```ts
+static fromRecordBatchReader(reader, opts): Promise<Scannable>
+```
+
+Build a Scannable from an Arrow `RecordBatchReader`. A reader can only
+be consumed once; `rescannable` defaults to `false`.
+
+The reader must already be opened (via `.open()`) so its `.schema` is
+populated. `RecordBatchReader.from(...)` returns an unopened reader.
+
+`opts.rescannable: true` is rejected because `RecordBatchReader` is a
+self-iterator (its `[Symbol.iterator]()` returns itself), and this
+constructor does not call `reader.reset()` between scans, so a second
+scan would always see an exhausted reader. For genuinely replayable
+sources, use
+`fromFactory(schema, () => openReader(), { rescannable: true })`,
+which mints a fresh reader on each scan.
+
+#### Parameters
+
+* **reader**: `RecordBatchReader`&lt;`any`&gt;
+
+* **opts**: [`ScannableOptions`](../interfaces/ScannableOptions.md) = `{}`
+
+#### Returns
+
+`Promise`&lt;[`Scannable`](Scannable.md)&gt;
+
+***
+
+### fromTable()
+
+```ts
+static fromTable(table, opts): Promise<Scannable>
+```
+
+Build a Scannable from an in-memory Arrow `Table`. Always rescannable;
+the table's batches are replayed on each scan.
+
+The table's row count is authoritative: `opts.numRows` must either be
+omitted or equal to `table.numRows`. `opts.rescannable` of `false` is
+rejected because in-memory Tables are always rescannable.
+
+#### Parameters
+
+* **table**: `Table`&lt;`any`&gt;
+
+* **opts**: [`ScannableOptions`](../interfaces/ScannableOptions.md) = `{}`
+
+#### Returns
+
+`Promise`&lt;[`Scannable`](Scannable.md)&gt;

docs/src/js/classes/Table.md

@@ -690,6 +690,74 @@ of the given query
 
 ***
 
+### setLsmWriteSpec()
+
+```ts
+abstract setLsmWriteSpec(spec): Promise<void>
+```
+
+Install an [LsmWriteSpec](../interfaces/LsmWriteSpec.md) on this table, selecting Lance's MemWAL
+LSM-style write path for future `mergeInsert` calls.
+
+`LsmWriteSpec` chooses one of three sharding strategies via `specType`:
+
+- `"bucket"` — hash-bucket writes by the single-column unenforced primary
+  key (`column` and `numBuckets` required).
+- `"identity"` — shard by the raw value of a scalar `column`.
+- `"unsharded"` — route every write to a single shard.
+
+All variants require the table to have an unenforced primary key
+([Table#setUnenforcedPrimaryKey](Table.md#setunenforcedprimarykey)); bucket sharding additionally
+requires it to be the single column being bucketed.
+
+#### Parameters
+
+* **spec**: [`LsmWriteSpec`](../interfaces/LsmWriteSpec.md)
+    The sharding spec to install.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+#### Example
+
+```ts
+await table.setUnenforcedPrimaryKey("id");
+await table.setLsmWriteSpec({
+  specType: "bucket",
+  column: "id",
+  numBuckets: 16,
+  maintainedIndexes: ["id_idx"],
+});
+```
+
+***
+
+### setUnenforcedPrimaryKey()
+
+```ts
+abstract setUnenforcedPrimaryKey(columns): Promise<void>
+```
+
+Set the unenforced primary key for this table to a single column.
+
+"Unenforced" means LanceDB does not check uniqueness on writes; the
+column is recorded in the schema as the primary key for use by features
+such as `merge_insert`. Only single-column primary keys are supported,
+and the key cannot be changed once set.
+
+#### Parameters
+
+* **columns**: `string` \| `string`[]
+    The primary key column. A one-element
+    array is also accepted; passing more than one column is rejected.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
 ### stats()
 
 ```ts
@@ -793,6 +861,23 @@ Return the table as an arrow table
 
 ***
 
+### unsetLsmWriteSpec()
+
+```ts
+abstract unsetLsmWriteSpec(): Promise<void>
+```
+
+Remove the [LsmWriteSpec](../interfaces/LsmWriteSpec.md) from this table, reverting to the standard
+`mergeInsert` write path.
+
+Errors if no spec is currently set.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
 ### update()
 
 #### update(opts)

docs/src/js/classes/VectorQuery.md

@@ -498,6 +498,30 @@ This is useful for pagination.
 
 ***
 
+### orderBy()
+
+```ts
+orderBy(ordering): this
+```
+
+Sort the results by the specified column(s).
+
+#### Parameters
+
+* **ordering**: [`ColumnOrdering`](../interfaces/ColumnOrdering.md) \| [`ColumnOrdering`](../interfaces/ColumnOrdering.md)[]
+
+#### Returns
+
+`this`
+
+This query builder.
+
+#### Inherited from
+
+`StandardQueryBase.orderBy`
+
+***
+
 ### outputSchema()
 
 ```ts

docs/src/js/functions/connectNamespace.md

@@ -0,0 +1,131 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / connectNamespace
+
+# Function: connectNamespace()
+
+## connectNamespace(implName, config, options)
+
+```ts
+function connectNamespace(
+   implName,
+   config,
+   options?): Promise<Connection>
+```
+
+Connect to a LanceDB database through a namespace.
+
+Unlike [connect](connect.md), which routes by URI scheme (local path vs.
+`db://` cloud), `connectNamespace` always returns a namespace-backed
+connection. The `implName` selects the namespace implementation:
+
+- `"dir"` — directory namespace, configured with [DirNamespaceConfig](../interfaces/DirNamespaceConfig.md).
+- `"rest"` — remote REST catalog, configured with [RestNamespaceConfig](../interfaces/RestNamespaceConfig.md).
+- Any other string — full module path for a custom implementation,
+  configured with a free-form string-keyed `properties` map.
+
+### Parameters
+
+* **implName**: `"dir"`
+
+* **config**: [`DirNamespaceConfig`](../interfaces/DirNamespaceConfig.md)
+
+* **options?**: `Partial`&lt;[`ConnectNamespaceOptions`](../interfaces/ConnectNamespaceOptions.md)&gt;
+
+### Returns
+
+`Promise`&lt;[`Connection`](../classes/Connection.md)&gt;
+
+### Examples
+
+```ts
+const db = await connectNamespace("dir", { root: "/path/to/db" });
+await db.createTable("users", [{ id: 1 }]);
+```
+
+```ts
+const db = await connectNamespace("rest", {
+  uri: "https://catalog.example.com",
+  headers: { "x-api-key": process.env.CATALOG_KEY ?? "" },
+});
+```
+
+```ts
+const db = await connectNamespace("my.custom.Namespace", {
+  endpoint: "...",
+});
+```
+
+## connectNamespace(implName, config, options)
+
+```ts
+function connectNamespace(
+   implName,
+   config,
+   options?): Promise<Connection>
+```
+
+Connect through the built-in REST namespace.
+
+Configured with [RestNamespaceConfig](../interfaces/RestNamespaceConfig.md). See the function-level
+documentation above for the full surface, examples, and how this
+relates to [connect](connect.md).
+
+### Parameters
+
+* **implName**: `"rest"`
+
+* **config**: [`RestNamespaceConfig`](../interfaces/RestNamespaceConfig.md)
+
+* **options?**: `Partial`&lt;[`ConnectNamespaceOptions`](../interfaces/ConnectNamespaceOptions.md)&gt;
+
+### Returns
+
+`Promise`&lt;[`Connection`](../classes/Connection.md)&gt;
+
+### Example
+
+```ts
+const db = await connectNamespace("rest", {
+  uri: "https://catalog.example.com",
+  headers: { "x-api-key": process.env.CATALOG_KEY ?? "" },
+});
+```
+
+## connectNamespace(implName, properties, options)
+
+```ts
+function connectNamespace(
+   implName,
+   properties,
+   options?): Promise<Connection>
+```
+
+Connect through a custom namespace implementation by full module path,
+configured with a free-form string-keyed `properties` map. Use the
+typed overloads above for the built-in `"dir"` and `"rest"` impls.
+
+See the function-level documentation above for examples and how this
+relates to [connect](connect.md).
+
+### Parameters
+
+* **implName**: `string`
+
+* **properties**: `Record`&lt;`string`, `string`&gt;
+
+* **options?**: `Partial`&lt;[`ConnectNamespaceOptions`](../interfaces/ConnectNamespaceOptions.md)&gt;
+
+### Returns
+
+`Promise`&lt;[`Connection`](../classes/Connection.md)&gt;
+
+### Example
+
+```ts
+const db = await connectNamespace("my.custom.Namespace", {
+  endpoint: "...",
+});
+```

docs/src/js/globals.md

@@ -32,6 +32,7 @@
 - [PhraseQuery](classes/PhraseQuery.md)
 - [Query](classes/Query.md)
 - [QueryBase](classes/QueryBase.md)
+- [Scannable](classes/Scannable.md)
 - [Session](classes/Session.md)
 - [StaticHeaderProvider](classes/StaticHeaderProvider.md)
 - [Table](classes/Table.md)
@@ -50,11 +51,19 @@
 - [AlterColumnsResult](interfaces/AlterColumnsResult.md)
 - [ClientConfig](interfaces/ClientConfig.md)
 - [ColumnAlteration](interfaces/ColumnAlteration.md)
+- [ColumnOrdering](interfaces/ColumnOrdering.md)
 - [CompactionStats](interfaces/CompactionStats.md)
+- [ConnectNamespaceOptions](interfaces/ConnectNamespaceOptions.md)
 - [ConnectionOptions](interfaces/ConnectionOptions.md)
+- [CreateNamespaceOptions](interfaces/CreateNamespaceOptions.md)
+- [CreateNamespaceResponse](interfaces/CreateNamespaceResponse.md)
 - [CreateTableOptions](interfaces/CreateTableOptions.md)
 - [DeleteResult](interfaces/DeleteResult.md)
+- [DescribeNamespaceResponse](interfaces/DescribeNamespaceResponse.md)
+- [DirNamespaceConfig](interfaces/DirNamespaceConfig.md)
 - [DropColumnsResult](interfaces/DropColumnsResult.md)
+- [DropNamespaceOptions](interfaces/DropNamespaceOptions.md)
+- [DropNamespaceResponse](interfaces/DropNamespaceResponse.md)
 - [ExecutableQuery](interfaces/ExecutableQuery.md)
 - [FragmentStatistics](interfaces/FragmentStatistics.md)
 - [FragmentSummaryStats](interfaces/FragmentSummaryStats.md)
@@ -69,13 +78,19 @@
 - [IvfFlatOptions](interfaces/IvfFlatOptions.md)
 - [IvfPqOptions](interfaces/IvfPqOptions.md)
 - [IvfRqOptions](interfaces/IvfRqOptions.md)
+- [ListNamespacesOptions](interfaces/ListNamespacesOptions.md)
+- [ListNamespacesResponse](interfaces/ListNamespacesResponse.md)
+- [LsmWriteSpec](interfaces/LsmWriteSpec.md)
 - [MergeResult](interfaces/MergeResult.md)
 - [OpenTableOptions](interfaces/OpenTableOptions.md)
 - [OptimizeOptions](interfaces/OptimizeOptions.md)
 - [OptimizeStats](interfaces/OptimizeStats.md)
 - [QueryExecutionOptions](interfaces/QueryExecutionOptions.md)
 - [RemovalStats](interfaces/RemovalStats.md)
+- [RenameTableOptions](interfaces/RenameTableOptions.md)
+- [RestNamespaceConfig](interfaces/RestNamespaceConfig.md)
 - [RetryConfig](interfaces/RetryConfig.md)
+- [ScannableOptions](interfaces/ScannableOptions.md)
 - [ShuffleOptions](interfaces/ShuffleOptions.md)
 - [SplitCalculatedOptions](interfaces/SplitCalculatedOptions.md)
 - [SplitHashOptions](interfaces/SplitHashOptions.md)
@@ -90,6 +105,7 @@
 - [UpdateResult](interfaces/UpdateResult.md)
 - [Version](interfaces/Version.md)
 - [WriteExecutionOptions](interfaces/WriteExecutionOptions.md)
+- [WriteProgress](interfaces/WriteProgress.md)
 
 ## Type Aliases
 
@@ -107,6 +123,7 @@
 
 - [RecordBatchIterator](functions/RecordBatchIterator.md)
 - [connect](functions/connect.md)
+- [connectNamespace](functions/connectNamespace.md)
 - [makeArrowTable](functions/makeArrowTable.md)
 - [packBits](functions/packBits.md)
 - [permutationBuilder](functions/permutationBuilder.md)

docs/src/js/interfaces/AddDataOptions.md

@@ -19,3 +19,39 @@ mode: "append" | "overwrite";
 If "append" (the default) then the new data will be added to the table
 
 If "overwrite" then the new data will replace the existing data in the table.
+
+***
+
+### progress()
+
+```ts
+progress: (progress) => void;
+```
+
+Optional callback invoked periodically with write progress.
+
+The callback is fired once per batch written and once more with
+`done: true` when the write completes. Calls are dispatched
+asynchronously to the JS event loop and never block the write — a slow
+callback will queue events rather than back-pressure the writer.
+
+Errors thrown from the callback are logged with `console.warn` and
+swallowed — they do not abort the write.
+
+#### Parameters
+
+* **progress**: [`WriteProgress`](WriteProgress.md)
+
+#### Returns
+
+`void`
+
+#### Example
+
+```ts
+await table.add(data, {
+  progress: (p) => {
+    console.log(`${p.outputRows}/${p.totalRows ?? "?"} rows`);
+  },
+});
+```

docs/src/js/interfaces/ColumnOrdering.md

@@ -0,0 +1,31 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / ColumnOrdering
+
+# Interface: ColumnOrdering
+
+## Properties
+
+### ascending?
+
+```ts
+optional ascending: boolean;
+```
+
+***
+
+### columnName
+
+```ts
+columnName: string;
+```
+
+***
+
+### nullsFirst?
+
+```ts
+optional nullsFirst: boolean;
+```

docs/src/js/interfaces/ConnectNamespaceOptions.md

@@ -0,0 +1,54 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / ConnectNamespaceOptions
+
+# Interface: ConnectNamespaceOptions
+
+## Properties
+
+### namespaceClientProperties?
+
+```ts
+optional namespaceClientProperties: Record<string, string>;
+```
+
+Extra properties for the backing namespace client.
+
+***
+
+### readConsistencyInterval?
+
+```ts
+optional readConsistencyInterval: number;
+```
+
+The interval, in seconds, at which to check for updates to the table
+from other processes. If None, then consistency is not checked. For
+performance reasons, this is the default. For strong consistency, set
+this to zero seconds. Then every read will check for updates from other
+processes. As a compromise, you can set this to a non-zero value for
+eventual consistency.
+
+***
+
+### session?
+
+```ts
+optional session: Session;
+```
+
+The session to use for this connection. Holds shared caches and other
+session-specific state.
+
+***
+
+### storageOptions?
+
+```ts
+optional storageOptions: Record<string, string>;
+```
+
+Configuration for object storage. The available options are described
+at https://docs.lancedb.com/storage/

docs/src/js/interfaces/CreateNamespaceOptions.md

@@ -0,0 +1,27 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / CreateNamespaceOptions
+
+# Interface: CreateNamespaceOptions
+
+## Properties
+
+### mode?
+
+```ts
+optional mode: "overwrite" | "create" | "exist_ok";
+```
+
+Creation mode.
+
+***
+
+### properties?
+
+```ts
+optional properties: Record<string, string>;
+```
+
+Properties to set on the new namespace.

docs/src/js/interfaces/CreateNamespaceResponse.md

@@ -0,0 +1,23 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / CreateNamespaceResponse
+
+# Interface: CreateNamespaceResponse
+
+## Properties
+
+### properties?
+
+```ts
+optional properties: Record<string, string>;
+```
+
+***
+
+### transactionId?
+
+```ts
+optional transactionId: string;
+```

docs/src/js/interfaces/DescribeNamespaceResponse.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DescribeNamespaceResponse
+
+# Interface: DescribeNamespaceResponse
+
+## Properties
+
+### properties?
+
+```ts
+optional properties: Record<string, string>;
+```

docs/src/js/interfaces/DirNamespaceConfig.md

@@ -0,0 +1,47 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DirNamespaceConfig
+
+# Interface: DirNamespaceConfig
+
+Configuration for the built-in directory namespace (`"dir"`).
+
+The directory namespace stores tables under a single root path (local
+filesystem or object storage URI). See
+[https://docs.lancedb.com/namespaces](https://docs.lancedb.com/namespaces) for the documented surface;
+less-common knobs live under [DirNamespaceConfig.extraProperties](DirNamespaceConfig.md#extraproperties).
+
+## Properties
+
+### extraProperties?
+
+```ts
+optional extraProperties: Record<string, string>;
+```
+
+Additional raw properties passed verbatim to the namespace
+implementation (e.g. `storage.*`, `credential_vendor.*`). Typed
+fields above take precedence on key collision.
+
+***
+
+### manifestEnabled?
+
+```ts
+optional manifestEnabled: boolean;
+```
+
+Whether to maintain a namespace manifest at the root. Required for
+child namespaces. Defaults to true on the impl side.
+
+***
+
+### root
+
+```ts
+root: string;
+```
+
+Root path or URI containing the LanceDB tables.

docs/src/js/interfaces/DropNamespaceOptions.md

@@ -0,0 +1,27 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DropNamespaceOptions
+
+# Interface: DropNamespaceOptions
+
+## Properties
+
+### behavior?
+
+```ts
+optional behavior: "restrict" | "cascade";
+```
+
+Refuse to drop if non-empty (restrict) or drop recursively (cascade).
+
+***
+
+### mode?
+
+```ts
+optional mode: "fail" | "skip";
+```
+
+Whether to skip if the namespace doesn't exist, or fail.

docs/src/js/interfaces/DropNamespaceResponse.md

@@ -0,0 +1,23 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DropNamespaceResponse
+
+# Interface: DropNamespaceResponse
+
+## Properties
+
+### properties?
+
+```ts
+optional properties: Record<string, string>;
+```
+
+***
+
+### transactionId?
+
+```ts
+optional transactionId: string[];
+```

docs/src/js/interfaces/ListNamespacesOptions.md

@@ -0,0 +1,27 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / ListNamespacesOptions
+
+# Interface: ListNamespacesOptions
+
+## Properties
+
+### limit?
+
+```ts
+optional limit: number;
+```
+
+An optional limit to the number of results to return.
+
+***
+
+### pageToken?
+
+```ts
+optional pageToken: string;
+```
+
+Token from a previous response for pagination.

docs/src/js/interfaces/ListNamespacesResponse.md

@@ -0,0 +1,23 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / ListNamespacesResponse
+
+# Interface: ListNamespacesResponse
+
+## Properties
+
+### namespaces
+
+```ts
+namespaces: string[];
+```
+
+***
+
+### pageToken?
+
+```ts
+optional pageToken: string;
+```

docs/src/js/interfaces/LsmWriteSpec.md

@@ -0,0 +1,64 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / LsmWriteSpec
+
+# Interface: LsmWriteSpec
+
+Specification selecting Lance's MemWAL LSM-style write path for
+`mergeInsert`.
+
+`specType` is `"bucket"`, `"identity"`, or `"unsharded"`. For `"bucket"`,
+`column` and `numBuckets` are required; for `"identity"`, `column` is
+required.
+
+## Properties
+
+### column?
+
+```ts
+optional column: string;
+```
+
+Bucket and identity variants: the sharding column.
+
+***
+
+### maintainedIndexes?
+
+```ts
+optional maintainedIndexes: string[];
+```
+
+Names of indexes the MemWAL should keep up to date during writes.
+
+***
+
+### numBuckets?
+
+```ts
+optional numBuckets: number;
+```
+
+Bucket variant: the number of buckets, in `[1, 1024]`.
+
+***
+
+### specType
+
+```ts
+specType: "bucket" | "identity" | "unsharded";
+```
+
+One of `"bucket"`, `"identity"`, or `"unsharded"`.
+
+***
+
+### writerConfigDefaults?
+
+```ts
+optional writerConfigDefaults: Record<string, string>;
+```
+
+Default `ShardWriter` configuration recorded in the MemWAL index.

docs/src/js/interfaces/RenameTableOptions.md

@@ -0,0 +1,29 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / RenameTableOptions
+
+# Interface: RenameTableOptions
+
+## Properties
+
+### namespacePath?
+
+```ts
+optional namespacePath: string[];
+```
+
+The namespace path of the table being renamed. Defaults to the root
+namespace (`[]`) when omitted.
+
+***
+
+### newNamespacePath?
+
+```ts
+optional newNamespacePath: string[];
+```
+
+The namespace path to move the table to as part of the rename. When
+omitted the table stays in `namespacePath`.

docs/src/js/interfaces/RestNamespaceConfig.md

@@ -0,0 +1,47 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / RestNamespaceConfig
+
+# Interface: RestNamespaceConfig
+
+Configuration for the built-in REST namespace (`"rest"`).
+
+The REST namespace talks to a remote catalog server over HTTP. See
+[https://docs.lancedb.com/namespaces](https://docs.lancedb.com/namespaces) for the documented surface;
+less-common knobs (TLS, metrics) live under
+[RestNamespaceConfig.extraProperties](RestNamespaceConfig.md#extraproperties).
+
+## Properties
+
+### extraProperties?
+
+```ts
+optional extraProperties: Record<string, string>;
+```
+
+Additional raw properties passed verbatim to the namespace
+implementation (e.g. `tls.*`, `ops_metrics_enabled`, `delimiter`).
+Typed fields above take precedence on key collision.
+
+***
+
+### headers?
+
+```ts
+optional headers: Record<string, string>;
+```
+
+HTTP headers forwarded with each request. Keys are passed through
+as-is (e.g. `"x-api-key"`, `"Authorization"`).
+
+***
+
+### uri
+
+```ts
+uri: string;
+```
+
+Catalog endpoint URL.

docs/src/js/interfaces/ScannableOptions.md

@@ -0,0 +1,29 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / ScannableOptions
+
+# Interface: ScannableOptions
+
+## Properties
+
+### numRows?
+
+```ts
+optional numRows: number;
+```
+
+Hint about the number of rows. Not validated against the stream.
+
+***
+
+### rescannable?
+
+```ts
+optional rescannable: boolean;
+```
+
+Whether the source can be scanned more than once. Defaults to `true` for
+`fromTable` / `fromFactory` and `false` for `fromIterable` /
+`fromRecordBatchReader`.

docs/src/js/interfaces/WriteProgress.md

@@ -0,0 +1,84 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / WriteProgress
+
+# Interface: WriteProgress
+
+Progress snapshot for a write operation, delivered to the `progress`
+callback passed to [Table.add](../classes/Table.md#add).
+
+## Properties
+
+### activeTasks
+
+```ts
+activeTasks: number;
+```
+
+Number of parallel write tasks currently in flight.
+
+***
+
+### done
+
+```ts
+done: boolean;
+```
+
+`true` for the final callback; `false` otherwise.
+
+***
+
+### elapsedSeconds
+
+```ts
+elapsedSeconds: number;
+```
+
+Wall-clock seconds since the write started.
+
+***
+
+### outputBytes
+
+```ts
+outputBytes: number;
+```
+
+Number of bytes written so far.
+
+***
+
+### outputRows
+
+```ts
+outputRows: number;
+```
+
+Number of rows written so far.
+
+***
+
+### totalRows?
+
+```ts
+optional totalRows: number;
+```
+
+Total rows expected, when the input source reports it.
+
+Always set on the final callback (the one with `done: true`), falling
+back to the actual number of rows written when the source could not
+report a row count up front.
+
+***
+
+### totalTasks
+
+```ts
+totalTasks: number;
+```
+
+Total number of parallel write tasks (the write parallelism).

docs/src/python/python.md

@@ -166,6 +166,12 @@ lists the indices that LanceDb supports.
 
 ::: lancedb.index.IvfFlat
 
+::: lancedb.index.IvfSq
+
+::: lancedb.index.IvfRq
+
+::: lancedb.index.HnswFlat
+
 ::: lancedb.table.IndexStatistics
 
 ## Querying (Asynchronous)

java/lancedb-core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
       <groupId>com.lancedb</groupId>
       <artifactId>lancedb-parent</artifactId>
-      <version>0.28.0-beta.11</version>
+      <version>0.29.1-beta.0</version>
       <relativePath>../pom.xml</relativePath>
     </parent>
 

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.28.0-beta.11</version>
+    <version>0.29.1-beta.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>7.0.0-beta.7</lance-core.version>
+        <lance-core.version>7.0.0-beta.13</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>

nodejs/AGENTS.md

@@ -3,11 +3,11 @@ The core Rust library is in the `../rust/lancedb` directory, the rust binding
 code is in the `src/` directory and the typescript bindings are in
 the `lancedb/` directory.
 
-Whenever you change the Rust code, you will need to recompile: `npm run build`.
+Whenever you change the Rust code, you will need to recompile: `pnpm build`.
 
 Common commands:
-* Build: `npm run build`
-* Lint: `npm run lint`
-* Fix lints: `npm run lint-fix`
-* Test: `npm test`
-* Run single test file: `npm test __test__/arrow.test.ts`
+* Build: `pnpm build`
+* Lint: `pnpm lint`
+* Fix lints: `pnpm lint-fix`
+* Test: `pnpm test`
+* Run single test file: `pnpm test __test__/arrow.test.ts`

nodejs/CONTRIBUTING.md

@@ -12,20 +12,22 @@ Typescript.
 * `src/`: Rust bindings source code
 * `lancedb/`: Typescript package source code
 * `__test__/`: Unit tests
-* `examples/`: An npm package with the examples shown in the documentation
+* `examples/`: A pnpm package with the examples shown in the documentation
 
 ## Development environment
 
 To set up your development environment, you will need to install the following:
 
-1. Node.js 14 or later
-2. Rust's package manager, Cargo. Use [rustup](https://rustup.rs/) to install.
-3. [protoc](https://grpc.io/docs/protoc-installation/) (Protocol Buffers compiler)
+1. Node.js 22 or later (required by pnpm 11)
+2. [pnpm](https://pnpm.io/installation) 11 or later (or run via `corepack enable`,
+   which uses the `packageManager` field in `package.json`)
+3. Rust's package manager, Cargo. Use [rustup](https://rustup.rs/) to install.
+4. [protoc](https://grpc.io/docs/protoc-installation/) (Protocol Buffers compiler)
 
 Initial setup:
 
 ```shell
-npm install
+pnpm install
 ```
 
 ### Commit Hooks
@@ -39,38 +41,38 @@ pre-commit install
 
 ## Development
 
-Most common development commands can be run using the npm scripts.
+Most common development commands can be run using the pnpm scripts.
 
 Build the package
 
 ```shell
-npm install
-npm run build
+pnpm install
+pnpm build
 ```
 
 Lint:
 
 ```shell
-npm run lint
+pnpm lint
 ```
 
 Format and fix lints:
 
 ```shell
-npm run lint-fix
+pnpm lint-fix
 ```
 
 Run tests:
 
 ```shell
-npm test
+pnpm test
 ```
 
 To run a single test:
 
 ```shell
 # Single file: table.test.ts
-npm test -- table.test.ts
+pnpm test -- table.test.ts
 # Single test: 'merge insert' in table.test.ts
-npm test -- table.test.ts --testNamePattern=merge\ insert
+pnpm test -- table.test.ts --testNamePattern=merge\ insert
 ```

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.28.0-beta.11"
+version = "0.29.1-beta.0"
 publish = false
 license.workspace = true
 description.workspace = true
@@ -22,6 +22,7 @@ arrow-schema.workspace = true
 env_logger.workspace = true
 futures.workspace = true
 lancedb = { path = "../rust/lancedb", default-features = false }
+lance-namespace.workspace = true
 napi = { version = "3.8.3", default-features = false, features = [
     "napi9",
     "async"

nodejs/__test__/connection.test.ts

@@ -4,7 +4,7 @@
 import { readdirSync } from "fs";
 import { Field, Float64, Schema } from "apache-arrow";
 import * as tmp from "tmp";
-import { Connection, Table, connect } from "../lancedb";
+import { Connection, Table, connect, connectNamespace } from "../lancedb";
 import { LocalTable } from "../lancedb/table";
 
 describe("when connecting", () => {
@@ -47,6 +47,14 @@ describe("given a connection", () => {
     await db.close();
     expect(db.isOpen()).toBe(false);
     await expect(db.tableNames()).rejects.toThrow("Connection is closed");
+    await expect(db.renameTable("a", "b")).rejects.toThrow(
+      "Connection is closed",
+    );
+  });
+
+  it("should report renameTable as unsupported on an OSS connection", async () => {
+    await db.createTable("a", [{ id: 1 }]);
+    await expect(db.renameTable("a", "b")).rejects.toThrow(/not supported/);
   });
   it("should be able to create a table from an object arg `createTable(options)`, or args `createTable(name, data, options)`", async () => {
     let tbl = await db.createTable("test", [{ id: 1 }, { id: 2 }]);
@@ -306,3 +314,186 @@ describe("clone table functionality", () => {
     ).rejects.toThrow("Deep clone is not yet implemented");
   });
 });
+
+describe("namespaces", () => {
+  let tmpDir: tmp.DirResult;
+  let db: Connection;
+
+  beforeEach(async () => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+    // The local DirectoryNamespace backend only supports child namespaces
+    // when manifest mode is enabled (see lance-namespace-impls/src/dir.rs).
+    db = await connect(tmpDir.name, {
+      // biome-ignore lint/style/useNamingConvention: opaque backend property key, must match Rust
+      namespaceClientProperties: { manifest_enabled: "true" },
+    });
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  it("should create and describe a namespace", async () => {
+    await db.createNamespace(["myns"]);
+    const desc = await db.describeNamespace(["myns"]);
+    expect(desc).toBeDefined();
+  });
+
+  it("should list namespaces created at the root", async () => {
+    await db.createNamespace(["alpha"]);
+    await db.createNamespace(["beta"]);
+    const list = await db.listNamespaces();
+    expect(list.namespaces).toEqual(expect.arrayContaining(["alpha", "beta"]));
+  });
+
+  it("should list child namespaces under a parent", async () => {
+    await db.createNamespace(["parent"]);
+    await db.createNamespace(["parent", "child"]);
+    const list = await db.listNamespaces(["parent"]);
+    expect(list.namespaces).toContain("child");
+  });
+
+  it("should drop a namespace", async () => {
+    await db.createNamespace(["ephemeral"]);
+    await db.dropNamespace(["ephemeral"]);
+    const list = await db.listNamespaces();
+    expect(list.namespaces).not.toContain("ephemeral");
+  });
+
+  it("should raise an error on any namespace op after close", async () => {
+    await db.close();
+    await expect(db.describeNamespace(["foo"])).rejects.toThrow(
+      "Connection is closed",
+    );
+    await expect(db.listNamespaces()).rejects.toThrow("Connection is closed");
+    await expect(db.createNamespace(["foo"])).rejects.toThrow(
+      "Connection is closed",
+    );
+    await expect(db.dropNamespace(["foo"])).rejects.toThrow(
+      "Connection is closed",
+    );
+  });
+
+  it("should raise an understandable error when describing a non-existent namespace", async () => {
+    await expect(db.describeNamespace(["does-not-exist"])).rejects.toThrow(
+      /not found/i,
+    );
+  });
+
+  it("should raise an error when creating a namespace that already exists", async () => {
+    await db.createNamespace(["dup"]);
+    await expect(db.createNamespace(["dup"])).rejects.toThrow();
+  });
+
+  it("should reject an unrecognized createNamespace mode with a clear error", async () => {
+    await expect(
+      // biome-ignore lint/suspicious/noExplicitAny: deliberately bypass TS to test runtime validation
+      db.createNamespace(["x"], { mode: "frobnicate" as any }),
+    ).rejects.toThrow(/Invalid mode 'frobnicate'/);
+  });
+
+  it("should reject an unrecognized dropNamespace mode with a clear error", async () => {
+    await db.createNamespace(["x"]);
+    await expect(
+      // biome-ignore lint/suspicious/noExplicitAny: deliberately bypass TS to test runtime validation
+      db.dropNamespace(["x"], { mode: "frobnicate" as any }),
+    ).rejects.toThrow(/Invalid mode 'frobnicate'/);
+  });
+
+  it("should reject an unrecognized dropNamespace behavior with a clear error", async () => {
+    await db.createNamespace(["x"]);
+    await expect(
+      // biome-ignore lint/suspicious/noExplicitAny: deliberately bypass TS to test runtime validation
+      db.dropNamespace(["x"], { behavior: "frobnicate" as any }),
+    ).rejects.toThrow(/Invalid behavior 'frobnicate'/);
+  });
+});
+
+describe("connectNamespace", () => {
+  let tmpDir: tmp.DirResult;
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  it("connects via the dir implementation and supports table ops", async () => {
+    const db = await connectNamespace("dir", { root: tmpDir.name });
+    await db.createTable("users", [{ id: 1 }, { id: 2 }]);
+    await expect(db.tableNames()).resolves.toContain("users");
+  });
+
+  it("throws a clear error when implName is empty", async () => {
+    await expect(connectNamespace("", {})).rejects.toThrow(
+      "implName must be a non-empty string",
+    );
+  });
+
+  it("throws when the namespace implementation is unknown", async () => {
+    await expect(connectNamespace("not-a-real-impl", {})).rejects.toThrow();
+  });
+
+  it("passes storage options through to the namespace", async () => {
+    const db = await connectNamespace(
+      "dir",
+      { root: tmpDir.name },
+      { storageOptions: { newTableDataStorageVersion: "stable" } },
+    );
+    await db.createTable("plumbing", [{ id: 1 }]);
+    await expect(db.tableNames()).resolves.toContain("plumbing");
+  });
+
+  it("supports child namespaces when manifestEnabled is true on the dir config", async () => {
+    const writer = await connectNamespace("dir", {
+      root: tmpDir.name,
+      manifestEnabled: true,
+    });
+    await writer.createNamespace(["analytics"]);
+    await writer.createTable("orders", [{ id: 1 }, { id: 2 }], ["analytics"]);
+    await writer.close();
+
+    const reader = await connectNamespace("dir", {
+      root: tmpDir.name,
+      manifestEnabled: true,
+    });
+    await expect(reader.tableNames(["analytics"])).resolves.toContain("orders");
+    const orders = await reader.openTable("orders", ["analytics"]);
+    await expect(orders.countRows()).resolves.toBe(2);
+  });
+
+  it("merges extraProperties into the dir config and is overridden by typed fields", async () => {
+    // Two observable assertions:
+    // - Typed `root` overrides extraProperties.root: createTable would fail
+    //   under the bogus path if the override didn't happen.
+    // - extraProperties.manifest_enabled="false" is honored end-to-end. Child
+    //   namespaces require manifest mode (default true), so explicitly
+    //   disabling it via extraProperties must make createNamespace reject. If
+    //   extraProperties pass-through were silently broken, the default would
+    //   let createNamespace succeed.
+    const db = await connectNamespace("dir", {
+      root: tmpDir.name,
+      extraProperties: {
+        root: "/should/be/overridden",
+        // biome-ignore lint/style/useNamingConvention: backend property key
+        manifest_enabled: "false",
+      },
+    });
+    await db.createTable("base", [{ id: 1 }]);
+    await expect(db.tableNames()).resolves.toContain("base");
+    await expect(db.createNamespace(["analytics"])).rejects.toThrow();
+  });
+
+  it("flows unknown top-level keys through when implName is dynamic (no silent drop)", async () => {
+    // Routes via the third overload because `impl` is `string`, not the
+    // literal `"dir"`. The dispatcher still notices the runtime value is
+    // "dir", but unknown keys like `manifest_enabled` must not be silently
+    // dropped during the conversion.
+    //
+    // Asserting a *negative* outcome (manifest disabled -> createNamespace
+    // rejects) is required for observability, since the backend default for
+    // `manifest_enabled` is true.
+    const impl: string = "dir";
+    const db = await connectNamespace(impl, {
+      root: tmpDir.name,
+      // biome-ignore lint/style/useNamingConvention: backend property key
+      manifest_enabled: "false",
+    });
+    await expect(db.createNamespace(["mixed"])).rejects.toThrow();
+  });
+});

nodejs/__test__/query.test.ts

@@ -109,3 +109,209 @@ describe("Query outputSchema", () => {
     expect(schema.fields.length).toBe(3);
   });
 });
+
+describe("Query orderBy", () => {
+  let tmpDir: tmp.DirResult;
+  let table: Table;
+
+  beforeEach(async () => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+    const db = await connect(tmpDir.name);
+
+    // Create table with numeric data for sorting
+    const schema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field("score", new Float32(), true),
+      new Field("name", new Utf8(), true),
+    ]);
+
+    const data = makeArrowTable(
+      [
+        { id: 1n, score: 3.5, name: "charlie" },
+        { id: 2n, score: 1.2, name: "alice" },
+        { id: 3n, score: 2.8, name: "bob" },
+        { id: 4n, score: 0.5, name: "david" },
+        { id: 5n, score: 4.1, name: "eve" },
+      ],
+      { schema },
+    );
+    table = await db.createTable("test", data);
+  });
+
+  afterEach(() => {
+    tmpDir.removeCallback();
+  });
+
+  it("should sort by single column ascending", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "score", ascending: true, nullsFirst: false })
+      .toArray();
+
+    expect(results.length).toBe(5);
+    // Verify ascending order
+    expect(results[0].score).toBeCloseTo(0.5, 0.001);
+    expect(results[1].score).toBeCloseTo(1.2, 0.001);
+    expect(results[2].score).toBeCloseTo(2.8, 0.001);
+    expect(results[3].score).toBeCloseTo(3.5, 0.001);
+    expect(results[4].score).toBeCloseTo(4.1, 0.001);
+  });
+
+  it("should sort by single column descending", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "score", ascending: false, nullsFirst: false })
+      .toArray();
+
+    expect(results.length).toBe(5);
+    // Verify descending order
+    expect(results[0].score).toBeCloseTo(4.1, 0.001);
+    expect(results[1].score).toBeCloseTo(3.5, 0.001);
+    expect(results[2].score).toBeCloseTo(2.8, 0.001);
+    expect(results[3].score).toBeCloseTo(1.2, 0.001);
+    expect(results[4].score).toBeCloseTo(0.5, 0.001);
+  });
+
+  it("should use ascending as default direction", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "score" })
+      .toArray();
+
+    expect(results.length).toBe(5);
+    // Verify ascending order (default)
+    expect(results[0].score).toBeCloseTo(0.5, 0.001);
+    expect(results[1].score).toBeCloseTo(1.2, 0.001);
+    expect(results[2].score).toBeCloseTo(2.8, 0.001);
+    expect(results[3].score).toBeCloseTo(3.5, 0.001);
+    expect(results[4].score).toBeCloseTo(4.1, 0.001);
+  });
+
+  it("should sort by string column", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "name" })
+      .toArray();
+
+    expect(results.length).toBe(5);
+    // Verify alphabetical order
+    expect(results[0].name).toBe("alice");
+    expect(results[1].name).toBe("bob");
+    expect(results[2].name).toBe("charlie");
+    expect(results[3].name).toBe("david");
+    expect(results[4].name).toBe("eve");
+  });
+
+  it("should support method chaining with where", async () => {
+    const results = await table
+      .query()
+      .where("score > 2.0")
+      .orderBy({ columnName: "score" })
+      .toArray();
+    expect(results.length).toBe(3);
+    // Verify filtered and sorted
+    expect(results[0].score).toBeCloseTo(2.8, 0.001);
+    expect(results[1].score).toBeCloseTo(3.5, 0.001);
+    expect(results[2].score).toBeCloseTo(4.1, 0.001);
+  });
+
+  it("should support method chaining with limit", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "score", ascending: false })
+      .limit(3)
+      .toArray();
+
+    expect(results.length).toBe(3);
+    // Verify top 3 in descending order
+    expect(results[0].score).toBeCloseTo(4.1, 0.001);
+    expect(results[1].score).toBeCloseTo(3.5, 0.001);
+    expect(results[2].score).toBeCloseTo(2.8, 0.001);
+  });
+
+  it("should support method chaining with offset", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "score" })
+      .offset(2)
+      .limit(2)
+      .toArray();
+
+    expect(results.length).toBe(2);
+    // Verify results skip first 2 and take next 2
+    expect(results[0].score).toBeCloseTo(2.8, 0.001);
+    expect(results[1].score).toBeCloseTo(3.5, 0.001);
+  });
+
+  it("should support method chaining with select", async () => {
+    const results = await table
+      .query()
+      .orderBy({ columnName: "name" })
+      .select(["name", "score"])
+      .toArray();
+
+    expect(results.length).toBe(5);
+    // Verify only selected columns are present
+    expect(Object.keys(results[0])).toEqual(["name", "score"]);
+    expect(Object.keys(results[4])).toEqual(["name", "score"]);
+    // Verify sorted by name
+    expect(results[0].name).toBe("alice");
+    expect(results[4].name).toBe("eve");
+  });
+
+  it("should support complex method chaining", async () => {
+    const results = await table
+      .query()
+      .where("score > 1.0")
+      .orderBy({ columnName: "score", ascending: false })
+      .limit(3)
+      .select(["id", "score", "name"])
+      .toArray();
+
+    expect(results.length).toBe(3);
+    // Verify filtered, sorted, limited, and projected
+    expect(results[0].score).toBeCloseTo(4.1, 0.001);
+    expect(results[1].score).toBeCloseTo(3.5, 0.001);
+    expect(results[2].score).toBeCloseTo(2.8, 0.001);
+    expect(Object.keys(results[0])).toEqual(["id", "score", "name"]);
+  });
+
+  it("should support multi-column ordering and null placement", async () => {
+    const schema = new Schema([
+      new Field("group", new Int64(), true),
+      new Field("score", new Float32(), true),
+      new Field("name", new Utf8(), true),
+    ]);
+
+    const data = makeArrowTable(
+      [
+        { group: 1n, score: null, name: "z" },
+        { group: 1n, score: 1.0, name: "b" },
+        { group: 1n, score: 1.0, name: "a" },
+        { group: 2n, score: 0.5, name: "c" },
+      ],
+      { schema },
+    );
+    const nullTable = await (await connect(tmpDir.name)).createTable(
+      "test_multi_order",
+      data,
+      { mode: "overwrite" },
+    );
+
+    const results = await nullTable
+      .query()
+      .orderBy([
+        { columnName: "group", ascending: true, nullsFirst: false },
+        { columnName: "score", ascending: true, nullsFirst: true },
+        { columnName: "name", ascending: true, nullsFirst: false },
+      ])
+      .toArray();
+
+    expect(results.map((r) => [r.group, r.score, r.name])).toEqual([
+      [1n, null, "z"],
+      [1n, 1.0, "a"],
+      [1n, 1.0, "b"],
+      [2n, 0.5, "c"],
+    ]);
+  });
+});

nodejs/__test__/remote.test.ts

@@ -617,4 +617,68 @@ describe("remote connection", () => {
       );
     });
   });
+
+  describe("renameTable", () => {
+    async function captureRenameRequest(
+      call: (db: Connection) => Promise<void>,
+    ): Promise<{ url: string; body: Record<string, unknown> }> {
+      let captured: { url: string; body: Record<string, unknown> } | undefined;
+      await withMockDatabase((req, res) => {
+        let raw = "";
+        req.on("data", (chunk) => {
+          raw += chunk;
+        });
+        req.on("end", () => {
+          captured = {
+            url: req.url ?? "",
+            body: raw ? JSON.parse(raw) : {},
+          };
+          res.writeHead(200, { "Content-Type": "application/json" }).end("");
+        });
+      }, call);
+      if (!captured) {
+        throw new Error("mock server never saw a request");
+      }
+      return captured;
+    }
+
+    it("sends rename request for a table in the root namespace", async () => {
+      const { url, body } = await captureRenameRequest(async (db) => {
+        await db.renameTable("table1", "table2");
+      });
+      expect(url).toBe("/v1/table/table1/rename/");
+      // biome-ignore lint/style/useNamingConvention: snake_case mandated by the server wire format
+      expect(body).toEqual({ new_table_name: "table2" });
+    });
+
+    it("omits new_namespace when only the current namespace is supplied", async () => {
+      // Safe-default check: passing namespacePath alone must not send
+      // `new_namespace`, so the server keeps the table in its current
+      // namespace instead of silently moving it to root.
+      const { url, body } = await captureRenameRequest(async (db) => {
+        await db.renameTable("table1", "table2", {
+          namespacePath: ["ns1"],
+        });
+      });
+      expect(url).toBe("/v1/table/ns1$table1/rename/");
+      // biome-ignore lint/style/useNamingConvention: snake_case mandated by the server wire format
+      expect(body).toEqual({ new_table_name: "table2" });
+    });
+
+    it("includes new_namespace in the body for a cross-namespace rename", async () => {
+      const { url, body } = await captureRenameRequest(async (db) => {
+        await db.renameTable("table1", "table2", {
+          namespacePath: ["ns1"],
+          newNamespacePath: ["ns2"],
+        });
+      });
+      expect(url).toBe("/v1/table/ns1$table1/rename/");
+      expect(body).toEqual({
+        // biome-ignore lint/style/useNamingConvention: snake_case mandated by the server wire format
+        new_table_name: "table2",
+        // biome-ignore lint/style/useNamingConvention: snake_case mandated by the server wire format
+        new_namespace: ["ns2"],
+      });
+    });
+  });
 });

nodejs/__test__/scannable.test.ts

@@ -0,0 +1,438 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+import {
+  Field,
+  Float16,
+  Int32,
+  type RecordBatch,
+  RecordBatchReader,
+  Schema,
+  tableToIPC,
+} from "apache-arrow";
+import { makeArrowTable, makeEmptyTable } from "../lancedb/arrow";
+import { Scannable } from "../lancedb/scannable";
+
+function makeTable() {
+  return makeArrowTable(
+    [
+      { id: 1, name: "a" },
+      { id: 2, name: "b" },
+      { id: 3, name: "c" },
+    ],
+    { vectorColumns: {} },
+  );
+}
+
+async function makeReader(): Promise<RecordBatchReader> {
+  // `RecordBatchReader.from()` returns an unopened reader; `.schema` is only
+  // populated after `.open()`. Opening sync readers is synchronous.
+  const reader = RecordBatchReader.from(tableToIPC(makeTable()));
+  return reader.open() as RecordBatchReader;
+}
+
+describe("Scannable", () => {
+  describe("fromTable", () => {
+    test("reflects schema, numRows, and defaults rescannable=true", async () => {
+      const table = makeTable();
+      const scannable = await Scannable.fromTable(table);
+
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.numRows).toBe(table.numRows);
+      expect(scannable.rescannable).toBe(true);
+    });
+
+    test("throws when opts.numRows does not match table.numRows", async () => {
+      await expect(
+        Scannable.fromTable(makeTable(), { numRows: 42 }),
+      ).rejects.toThrow(/does not match table\.numRows/);
+    });
+
+    test("throws when opts.rescannable is false", async () => {
+      await expect(
+        Scannable.fromTable(makeTable(), { rescannable: false }),
+      ).rejects.toThrow(/always rescannable/);
+    });
+  });
+
+  describe("fromRecordBatchReader", () => {
+    test("reflects schema and defaults numRows=null, rescannable=false", async () => {
+      const reader = await makeReader();
+      const scannable = await Scannable.fromRecordBatchReader(reader);
+
+      expect(scannable.schema).toBe(reader.schema);
+      expect(scannable.numRows).toBeNull();
+      expect(scannable.rescannable).toBe(false);
+    });
+
+    test("honors numRows override", async () => {
+      const scannable = await Scannable.fromRecordBatchReader(
+        await makeReader(),
+        { numRows: 3 },
+      );
+
+      expect(scannable.numRows).toBe(3);
+      expect(scannable.rescannable).toBe(false);
+    });
+
+    test("rescannable: false explicit does not throw", async () => {
+      const reader = await makeReader();
+      const scannable = await Scannable.fromRecordBatchReader(reader, {
+        rescannable: false,
+      });
+      expect(scannable.rescannable).toBe(false);
+    });
+
+    test("throws when opts.rescannable is true", async () => {
+      const reader = await makeReader();
+      await expect(
+        Scannable.fromRecordBatchReader(reader, { rescannable: true }),
+      ).rejects.toThrow(/does not accept rescannable/);
+    });
+
+    test("throws when opts.rescannable is true even alongside numRows", async () => {
+      const reader = await makeReader();
+      await expect(
+        Scannable.fromRecordBatchReader(reader, {
+          numRows: 3,
+          rescannable: true,
+        }),
+      ).rejects.toThrow(/does not accept rescannable/);
+    });
+  });
+
+  describe("fromIterable", () => {
+    test("accepts a sync iterable of batches", async () => {
+      const table = makeTable();
+      const scannable = await Scannable.fromIterable(
+        table.schema,
+        table.batches,
+      );
+
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.numRows).toBeNull();
+      expect(scannable.rescannable).toBe(false);
+    });
+
+    test("accepts an async iterable of batches", async () => {
+      const table = makeTable();
+      async function* generator(): AsyncGenerator<RecordBatch> {
+        for (const batch of table.batches) {
+          yield batch;
+        }
+      }
+
+      const scannable = await Scannable.fromIterable(table.schema, generator());
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.rescannable).toBe(false);
+    });
+
+    describe("rescannable: true detection", () => {
+      // Replayable inputs: [Symbol.iterator]() / [Symbol.asyncIterator]()
+      // returns a fresh iterator each call. Must NOT throw.
+
+      test("Array passes (fresh ArrayIterator each call)", async () => {
+        const table = makeTable();
+        const scannable = await Scannable.fromIterable(
+          table.schema,
+          table.batches,
+          { rescannable: true },
+        );
+        expect(scannable.rescannable).toBe(true);
+      });
+
+      test("Set passes (fresh SetIterator each call)", async () => {
+        const table = makeTable();
+        const set = new Set<RecordBatch>(table.batches);
+        const scannable = await Scannable.fromIterable(table.schema, set, {
+          rescannable: true,
+        });
+        expect(scannable.rescannable).toBe(true);
+      });
+
+      test("custom Iterable returning a fresh iterator passes", async () => {
+        const table = makeTable();
+        const replayable: Iterable<RecordBatch> = {
+          [Symbol.iterator]() {
+            return table.batches[Symbol.iterator]();
+          },
+        };
+        const scannable = await Scannable.fromIterable(
+          table.schema,
+          replayable,
+          { rescannable: true },
+        );
+        expect(scannable.rescannable).toBe(true);
+      });
+
+      test("object with generator method passes (fresh generator each call)", async () => {
+        const table = makeTable();
+        const replayable: Iterable<RecordBatch> = {
+          *[Symbol.iterator]() {
+            for (const batch of table.batches) yield batch;
+          },
+        };
+        const scannable = await Scannable.fromIterable(
+          table.schema,
+          replayable,
+          { rescannable: true },
+        );
+        expect(scannable.rescannable).toBe(true);
+      });
+
+      test("empty Array passes (replayable degenerate case)", async () => {
+        const schema = makeTable().schema;
+        const scannable = await Scannable.fromIterable(
+          schema,
+          [] as RecordBatch[],
+          { rescannable: true },
+        );
+        expect(scannable.rescannable).toBe(true);
+      });
+
+      // One-shot inputs: [Symbol.iterator]() / [Symbol.asyncIterator]()
+      // returns the same object, or the input is already-an-iterator.
+      // Must throw with a /one-shot/ message.
+
+      test("sync generator throws", async () => {
+        const table = makeTable();
+        function* generator(): Generator<RecordBatch> {
+          for (const batch of table.batches) yield batch;
+        }
+        await expect(
+          Scannable.fromIterable(table.schema, generator(), {
+            rescannable: true,
+          }),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      test("async generator throws", async () => {
+        const table = makeTable();
+        async function* generator(): AsyncGenerator<RecordBatch> {
+          for (const batch of table.batches) yield batch;
+        }
+        await expect(
+          Scannable.fromIterable(table.schema, generator(), {
+            rescannable: true,
+          }),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      test("empty generator throws (one-shot degenerate case)", async () => {
+        const schema = makeTable().schema;
+        function* generator(): Generator<RecordBatch> {
+          // intentionally empty; yields nothing.
+        }
+        await expect(
+          Scannable.fromIterable(schema, generator(), { rescannable: true }),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      test("custom self-iterator throws", async () => {
+        const table = makeTable();
+        const batches = table.batches;
+        let i = 0;
+        const oneShot: Iterable<RecordBatch> & Iterator<RecordBatch> = {
+          [Symbol.iterator]() {
+            return this;
+          },
+          next() {
+            if (i >= batches.length) {
+              return { done: true, value: undefined };
+            }
+            return { done: false, value: batches[i++] };
+          },
+        };
+        await expect(
+          Scannable.fromIterable(table.schema, oneShot, { rescannable: true }),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      test("Array.values() (IterableIterator) throws", async () => {
+        const table = makeTable();
+        const iter = table.batches.values();
+        await expect(
+          Scannable.fromIterable(table.schema, iter, { rescannable: true }),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      test("raw iterator (only `.next`) throws", async () => {
+        const table = makeTable();
+        const batches = table.batches;
+        let i = 0;
+        const rawIter = {
+          next(): IteratorResult<RecordBatch> {
+            if (i >= batches.length) {
+              return { done: true, value: undefined };
+            }
+            return { done: false, value: batches[i++] };
+          },
+        };
+        await expect(
+          Scannable.fromIterable(
+            table.schema,
+            rawIter as unknown as Iterable<RecordBatch>,
+            { rescannable: true },
+          ),
+        ).rejects.toThrow(/one-shot/);
+      });
+
+      // Edge: null/undefined must not crash the detection helper. The
+      // null check belongs to `normalizeIterator` and only fires when a
+      // scan starts.
+
+      test("null input does not crash detection at construction", async () => {
+        const schema = makeTable().schema;
+        await expect(
+          Scannable.fromIterable(
+            schema,
+            null as unknown as Iterable<RecordBatch>,
+            {
+              rescannable: true,
+            },
+          ),
+        ).resolves.toBeDefined();
+      });
+
+      test("undefined input does not crash detection at construction", async () => {
+        const schema = makeTable().schema;
+        await expect(
+          Scannable.fromIterable(
+            schema,
+            undefined as unknown as Iterable<RecordBatch>,
+            { rescannable: true },
+          ),
+        ).resolves.toBeDefined();
+      });
+
+      // Default (rescannable omitted) skips the check entirely, so even
+      // pathological inputs construct without throwing here.
+
+      test("rescannable omitted skips detection entirely (generator passes)", async () => {
+        const table = makeTable();
+        function* generator(): Generator<RecordBatch> {
+          for (const batch of table.batches) yield batch;
+        }
+        const scannable = await Scannable.fromIterable(
+          table.schema,
+          generator(),
+        );
+        expect(scannable.rescannable).toBe(false);
+      });
+
+      test("rescannable: false explicit skips detection entirely (generator passes)", async () => {
+        const table = makeTable();
+        function* generator(): Generator<RecordBatch> {
+          for (const batch of table.batches) yield batch;
+        }
+        const scannable = await Scannable.fromIterable(
+          table.schema,
+          generator(),
+          { rescannable: false },
+        );
+        expect(scannable.rescannable).toBe(false);
+      });
+    });
+  });
+
+  describe("fromFactory", () => {
+    test("defaults rescannable=true and does not invoke the factory eagerly", async () => {
+      const table = makeTable();
+      const factory = jest.fn(() => table.batches);
+
+      const scannable = await Scannable.fromFactory(table.schema, factory);
+
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.rescannable).toBe(true);
+      expect(factory).not.toHaveBeenCalled();
+    });
+
+    test("honors rescannable and numRows overrides", async () => {
+      const table = makeTable();
+      const scannable = await Scannable.fromFactory(
+        table.schema,
+        () => table.batches,
+        { numRows: 7, rescannable: false },
+      );
+
+      expect(scannable.numRows).toBe(7);
+      expect(scannable.rescannable).toBe(false);
+    });
+  });
+
+  describe("validation", () => {
+    test("throws when numRows is negative", async () => {
+      await expect(
+        Scannable.fromFactory(makeTable().schema, () => [], { numRows: -1 }),
+      ).rejects.toThrow(/non-negative/);
+    });
+
+    test("throws when numRows is not an integer", async () => {
+      await expect(
+        Scannable.fromFactory(makeTable().schema, () => [], { numRows: 3.5 }),
+      ).rejects.toThrow(/integer/);
+    });
+  });
+
+  describe("native handle", () => {
+    test("exposes a native handle via inner", async () => {
+      const scannable = await Scannable.fromTable(makeTable());
+      expect(scannable.inner).toBeDefined();
+      expect(typeof scannable.inner).toBe("object");
+      expect(scannable.inner).not.toBeNull();
+    });
+  });
+
+  // Schema-variety construction tests. Each asserts that construction
+  // succeeds against a richer Arrow schema, which transitively exercises
+  // schema serialization and the Rust-side `ipc_file_to_schema` for types
+  // beyond flat primitives.
+  describe("schema variety", () => {
+    test("accepts an empty table", async () => {
+      const schema = new Schema([new Field("id", new Int32(), true)]);
+      const table = makeEmptyTable(schema);
+      const scannable = await Scannable.fromTable(table);
+
+      expect(scannable.numRows).toBe(0);
+      expect(scannable.schema).toBe(table.schema);
+    });
+
+    test("accepts nested struct and list columns", async () => {
+      const table = makeArrowTable(
+        [
+          { id: 1, point: { x: 0, y: 0 }, tags: ["a", "b"] },
+          { id: 2, point: { x: 1, y: 2 }, tags: ["c"] },
+        ],
+        { vectorColumns: {} },
+      );
+      const scannable = await Scannable.fromTable(table);
+
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.numRows).toBe(2);
+    });
+
+    test("accepts a FixedSizeList (vector) column", async () => {
+      const table = makeArrowTable(
+        [
+          { id: 1, vec: [1, 2, 3] },
+          { id: 2, vec: [4, 5, 6] },
+        ],
+        { vectorColumns: { vec: { type: new Float16() } } },
+      );
+      const scannable = await Scannable.fromTable(table);
+
+      expect(scannable.schema).toBe(table.schema);
+      expect(scannable.numRows).toBe(2);
+    });
+
+    test("accepts a table with many columns", async () => {
+      const row: Record<string, number> = {};
+      for (let i = 0; i < 50; i++) row[`c${i}`] = i;
+      const table = makeArrowTable([row, row], { vectorColumns: {} });
+      const scannable = await Scannable.fromTable(table);
+
+      expect(scannable.schema.fields.length).toBe(50);
+      expect(scannable.numRows).toBe(2);
+    });
+  });
+});

nodejs/__test__/table.test.ts

@@ -115,6 +115,48 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       await expect(table.countRows()).resolves.toBe(1);
     });
 
+    it("should invoke the progress callback", async () => {
+      const events: import("../lancedb").WriteProgress[] = [];
+      await table.add([{ id: 1 }, { id: 2 }, { id: 3 }], {
+        progress: (p) => events.push(p),
+      });
+
+      expect(events.length).toBeGreaterThan(0);
+      const last = events[events.length - 1];
+      expect(last.done).toBe(true);
+      // Earlier callbacks must have done=false.
+      for (const ev of events.slice(0, -1)) {
+        expect(ev.done).toBe(false);
+      }
+      // outputRows reflects the rows added in this call, not table size.
+      expect(last.outputRows).toBe(3);
+      // The input source (an array) reports a row count, so totalRows is set.
+      expect(last.totalRows).toBe(3);
+      // outputRows is monotonic.
+      for (let i = 1; i < events.length; i++) {
+        expect(events[i].outputRows).toBeGreaterThanOrEqual(
+          events[i - 1].outputRows,
+        );
+      }
+    });
+
+    it("should swallow errors thrown from the progress callback", async () => {
+      const warn = jest
+        .spyOn(console, "warn")
+        .mockImplementation(() => undefined);
+      try {
+        const res = await table.add([{ id: 1 }, { id: 2 }], {
+          progress: () => {
+            throw new Error("callback bomb");
+          },
+        });
+        expect(res.version).toBeGreaterThan(0);
+        expect(warn).toHaveBeenCalled();
+      } finally {
+        warn.mockRestore();
+      }
+    });
+
     it("should let me close the table", async () => {
       expect(table.isOpen()).toBe(true);
       table.close();
@@ -2348,3 +2390,130 @@ describe("when creating a table with Float32Array vectors", () => {
     expect((fsl.children[0].type as Float32).precision).toBe(1);
   });
 });
+
+describe("setUnenforcedPrimaryKey", () => {
+  let tmpDir: tmp.DirResult;
+
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  it("sets a single-column primary key (string or one-element array)", async () => {
+    const conn = await connect(tmpDir.name);
+    const schema = new arrow.Schema([
+      new arrow.Field("id", new arrow.Int64(), false),
+    ]);
+    const t1 = await conn.createEmptyTable("t1", schema);
+    await t1.setUnenforcedPrimaryKey("id");
+
+    const t2 = await conn.createEmptyTable("t2", schema);
+    await t2.setUnenforcedPrimaryKey(["id"]);
+  });
+
+  it("rejects a compound primary key", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await conn.createEmptyTable(
+      "t",
+      new arrow.Schema([
+        new arrow.Field("id", new arrow.Int64(), false),
+        new arrow.Field("name", new arrow.Utf8(), false),
+      ]),
+    );
+    await expect(
+      table.setUnenforcedPrimaryKey(["id", "name"]),
+    ).rejects.toThrow();
+  });
+
+  it("rejects changing the primary key once set", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await conn.createEmptyTable(
+      "t",
+      new arrow.Schema([
+        new arrow.Field("id", new arrow.Int64(), false),
+        new arrow.Field("name", new arrow.Utf8(), false),
+      ]),
+    );
+    await table.setUnenforcedPrimaryKey("id");
+    await expect(table.setUnenforcedPrimaryKey("name")).rejects.toThrow();
+    await expect(table.setUnenforcedPrimaryKey("id")).rejects.toThrow();
+  });
+});
+
+describe("setLsmWriteSpec / unsetLsmWriteSpec", () => {
+  let tmpDir: tmp.DirResult;
+
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  async function makeTable(conn: Connection): Promise<Table> {
+    return await conn.createEmptyTable(
+      "t",
+      new arrow.Schema([new arrow.Field("id", new arrow.Int64(), false)]),
+    );
+  }
+
+  it("installs and removes a bucket spec", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await makeTable(conn);
+
+    await table.setUnenforcedPrimaryKey("id");
+    await table.setLsmWriteSpec({
+      specType: "bucket",
+      column: "id",
+      numBuckets: 4,
+    });
+    await table.unsetLsmWriteSpec();
+    // A second unset errors — there is no spec left to remove.
+    await expect(table.unsetLsmWriteSpec()).rejects.toThrow();
+    // A fresh spec can be installed after unset.
+    await table.setLsmWriteSpec({
+      specType: "bucket",
+      column: "id",
+      numBuckets: 8,
+    });
+  });
+
+  it("installs an unsharded spec", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await makeTable(conn);
+
+    await table.setUnenforcedPrimaryKey("id");
+    await table.setLsmWriteSpec({ specType: "unsharded" });
+    await table.unsetLsmWriteSpec();
+  });
+
+  it("installs an identity spec", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await makeTable(conn);
+
+    await table.setUnenforcedPrimaryKey("id");
+    await table.setLsmWriteSpec({ specType: "identity", column: "id" });
+    await table.unsetLsmWriteSpec();
+  });
+
+  it("rejects an invalid spec", async () => {
+    const conn = await connect(tmpDir.name);
+    const table = await makeTable(conn);
+
+    await table.setUnenforcedPrimaryKey("id");
+    // num_buckets out of range.
+    await expect(
+      table.setLsmWriteSpec({
+        specType: "bucket",
+        column: "id",
+        numBuckets: 0,
+      }),
+    ).rejects.toThrow();
+    // Column mismatch.
+    await expect(
+      table.setLsmWriteSpec({
+        specType: "bucket",
+        column: "missing",
+        numBuckets: 4,
+      }),
+    ).rejects.toThrow();
+  });
+});

nodejs/examples/filtering.test.ts

@@ -38,5 +38,14 @@ test("filtering examples", async () => {
     // --8<-- [start:sql_search]
     await tbl.query().where("id = 10").limit(10).toArray();
     // --8<-- [end:sql_search]
+
+    // --8<-- [start:orderby_search]
+    await tbl
+      .query()
+      .where("id > 10")
+      .orderBy({ columnName: "id", ascending: false })
+      .limit(5)
+      .toArray();
+    // --8<-- [end:orderby_search]
   });
 });

nodejs/examples/package.json

@@ -11,16 +11,17 @@
     "test": "node --experimental-vm-modules node_modules/.bin/jest --testEnvironment jest-environment-node-single-context --verbose",
     "lint": "biome check *.ts && biome format *.ts",
     "lint-ci": "biome ci .",
-    "lint-fix": "biome check --write *.ts && npm run format",
+    "lint-fix": "biome check --write *.ts && pnpm format",
     "format": "biome format --write *.ts"
   },
   "author": "Lance Devs",
   "license": "Apache-2.0",
+  "packageManager": "pnpm@11.1.1",
   "dependencies": {
-    "@huggingface/transformers": "^3.0.2",
+    "@huggingface/transformers": "3.0.2",
     "@lancedb/lancedb": "file:../dist",
-    "openai": "^4.29.2",
-    "sharp": "^0.33.5"
+    "openai": "4.29.2",
+    "sharp": "0.33.5"
   },
   "devDependencies": {
     "@biomejs/biome": "^1.7.3",

nodejs/examples/pnpm-workspace.yaml

@@ -0,0 +1,13 @@
+# Block resolution of versions less than 24h old (Shai-Hulud window).
+# This is the pnpm 11 default but pinned here so it's visible to
+# reviewers and survives a future pnpm major flipping the default.
+minimumReleaseAge: 1440
+
+# Fail install if a transitive dep tries to run an unapproved script.
+strictDepBuilds: true
+
+allowBuilds:
+  '@biomejs/biome': true
+  onnxruntime-node: true
+  protobufjs: true
+  sharp: true

nodejs/lancedb/arrow.ts

@@ -1291,6 +1291,18 @@ export async function fromRecordBatchToBuffer(
   return Buffer.from(await writer.toUint8Array());
 }
 
+/**
+ * Create a buffer containing a single record batch using the Arrow IPC Stream
+ * serialization. Each call produces a self-contained Stream message (schema +
+ * batch + EOS) suitable for incremental decode by `arrow_ipc::reader::StreamReader`.
+ */
+export async function fromRecordBatchToStreamBuffer(
+  batch: RecordBatch,
+): Promise<Buffer> {
+  const writer = RecordBatchStreamWriter.writeAll([batch]);
+  return Buffer.from(await writer.toUint8Array());
+}
+
 /**
  * Serialize an Arrow Table into a buffer using the Arrow IPC Stream serialization
  *

nodejs/lancedb/connection.ts

@@ -16,6 +16,18 @@ import {
 } from "./arrow";
 import { EmbeddingFunctionConfig, getRegistry } from "./embedding/registry";
 import { Connection as LanceDbConnection } from "./native";
+import type {
+  CreateNamespaceResponse,
+  DescribeNamespaceResponse,
+  DropNamespaceResponse,
+  ListNamespacesResponse,
+} from "./native";
+export type {
+  CreateNamespaceResponse,
+  DescribeNamespaceResponse,
+  DropNamespaceResponse,
+  ListNamespacesResponse,
+};
 import { sanitizeTable } from "./sanitize";
 import { LocalTable, Table } from "./table";
 
@@ -110,6 +122,41 @@ export interface TableNamesOptions {
   /** An optional limit to the number of results to return. */
   limit?: number;
 }
+
+export interface ListNamespacesOptions {
+  /** Token from a previous response for pagination. */
+  pageToken?: string;
+  /** An optional limit to the number of results to return. */
+  limit?: number;
+}
+
+export interface CreateNamespaceOptions {
+  /** Creation mode. */
+  mode?: "create" | "exist_ok" | "overwrite";
+  /** Properties to set on the new namespace. */
+  properties?: Record<string, string>;
+}
+
+export interface DropNamespaceOptions {
+  /** Whether to skip if the namespace doesn't exist, or fail. */
+  mode?: "skip" | "fail";
+  /** Refuse to drop if non-empty (restrict) or drop recursively (cascade). */
+  behavior?: "restrict" | "cascade";
+}
+
+export interface RenameTableOptions {
+  /**
+   * The namespace path of the table being renamed. Defaults to the root
+   * namespace (`[]`) when omitted.
+   */
+  namespacePath?: string[];
+  /**
+   * The namespace path to move the table to as part of the rename. When
+   * omitted the table stays in `namespacePath`.
+   */
+  newNamespacePath?: string[];
+}
+
 /**
  * A LanceDB Connection that allows you to open tables and create new ones.
  *
@@ -268,6 +315,69 @@ export abstract class Connection {
    */
   abstract dropAllTables(namespacePath?: string[]): Promise<void>;
 
+  /**
+   * Describe a namespace, returning its properties.
+   *
+   * @param {string[]} namespacePath - The namespace path to describe, in
+   *   parent → child order, e.g. `["analytics", "sales"]`.
+   * @returns {Promise<DescribeNamespaceResponse>} The namespace's properties
+   *   (may be undefined if the namespace has none).
+   */
+  abstract describeNamespace(
+    namespacePath: string[],
+  ): Promise<DescribeNamespaceResponse>;
+
+  /**
+   * List the immediate child namespaces under the given parent.
+   *
+   * Results may be paginated. To retrieve subsequent pages, pass the
+   * `pageToken` returned by a previous call.
+   *
+   * @param {string[]} namespacePath - The parent namespace path. Defaults
+   *   to the root namespace if omitted.
+   * @param {Partial<ListNamespacesOptions>} options - Pagination options
+   *   (`pageToken`, `limit`).
+   * @returns {Promise<ListNamespacesResponse>} Child namespace names and
+   *   an optional token for fetching the next page.
+   */
+  abstract listNamespaces(
+    namespacePath?: string[],
+    options?: Partial<ListNamespacesOptions>,
+  ): Promise<ListNamespacesResponse>;
+
+  /**
+   * Create a new namespace at the given path.
+   *
+   * @param {string[]} namespacePath - The namespace path to create.
+   * @param {Partial<CreateNamespaceOptions>} options - Creation `mode`
+   *   ("create" | "exist_ok" | "overwrite") and optional `properties`
+   *   to attach to the namespace.
+   * @returns {Promise<CreateNamespaceResponse>} The properties of the
+   *   created namespace and an optional transaction id.
+   */
+  abstract createNamespace(
+    namespacePath: string[],
+    options?: Partial<CreateNamespaceOptions>,
+  ): Promise<CreateNamespaceResponse>;
+
+  /**
+   * Drop a namespace.
+   *
+   * Use `behavior: "cascade"` to also drop everything contained in the
+   * namespace (sub-namespaces and tables). The default `"restrict"`
+   * behavior refuses to drop a non-empty namespace.
+   *
+   * @param {string[]} namespacePath - The namespace path to drop.
+   * @param {Partial<DropNamespaceOptions>} options - `mode` ("skip" | "fail"
+   *   for missing-namespace handling) and `behavior` ("restrict" | "cascade").
+   * @returns {Promise<DropNamespaceResponse>} Any properties returned by
+   *   the server and an optional transaction id.
+   */
+  abstract dropNamespace(
+    namespacePath: string[],
+    options?: Partial<DropNamespaceOptions>,
+  ): Promise<DropNamespaceResponse>;
+
   /**
    * Clone a table from a source table.
    *
@@ -294,6 +404,24 @@ export abstract class Connection {
       isShallow?: boolean;
     },
   ): Promise<Table>;
+
+  /**
+   * Rename a table.
+   *
+   * Currently only supported by LanceDB Cloud. Local OSS connections and
+   * namespace-backed connections (via {@link connectNamespace}) reject with
+   * a "not supported" error.
+   *
+   * @param {string} currentName - The current name of the table.
+   * @param {string} newName - The new name for the table.
+   * @param {RenameTableOptions} options - Optional namespace paths. When
+   *   `newNamespacePath` is omitted the table stays in `namespacePath`.
+   */
+  abstract renameTable(
+    currentName: string,
+    newName: string,
+    options?: RenameTableOptions,
+  ): Promise<void>;
 }
 
 /** @hideconstructor */
@@ -515,6 +643,58 @@ export class LocalConnection extends Connection {
   async dropAllTables(namespacePath?: string[]): Promise<void> {
     return this.inner.dropAllTables(namespacePath ?? []);
   }
+
+  describeNamespace(
+    namespacePath: string[],
+  ): Promise<DescribeNamespaceResponse> {
+    return this.inner.describeNamespace(namespacePath);
+  }
+
+  listNamespaces(
+    namespacePath?: string[],
+    options?: Partial<ListNamespacesOptions>,
+  ): Promise<ListNamespacesResponse> {
+    return this.inner.listNamespaces(
+      namespacePath ?? [],
+      options?.pageToken,
+      options?.limit,
+    );
+  }
+
+  createNamespace(
+    namespacePath: string[],
+    options?: Partial<CreateNamespaceOptions>,
+  ): Promise<CreateNamespaceResponse> {
+    return this.inner.createNamespace(
+      namespacePath,
+      options?.mode,
+      options?.properties,
+    );
+  }
+
+  dropNamespace(
+    namespacePath: string[],
+    options?: Partial<DropNamespaceOptions>,
+  ): Promise<DropNamespaceResponse> {
+    return this.inner.dropNamespace(
+      namespacePath,
+      options?.mode,
+      options?.behavior,
+    );
+  }
+
+  async renameTable(
+    currentName: string,
+    newName: string,
+    options?: RenameTableOptions,
+  ): Promise<void> {
+    return this.inner.renameTable(
+      currentName,
+      newName,
+      options?.namespacePath ?? [],
+      options?.newNamespacePath,
+    );
+  }
 }
 
 /**

nodejs/lancedb/index.ts

@@ -8,6 +8,7 @@ import {
 } from "./connection";
 
 import {
+  ConnectNamespaceOptions,
   ConnectionOptions,
   Connection as LanceDbConnection,
   JsHeaderProvider as NativeJsHeaderProvider,
@@ -22,6 +23,7 @@ export { JsHeaderProvider as NativeJsHeaderProvider } from "./native.js";
 export {
   AddColumnsSql,
   ConnectionOptions,
+  ConnectNamespaceOptions,
   IndexStatistics,
   IndexConfig,
   ClientConfig,
@@ -62,6 +64,14 @@ export {
   CreateTableOptions,
   TableNamesOptions,
   OpenTableOptions,
+  ListNamespacesOptions,
+  CreateNamespaceOptions,
+  DropNamespaceOptions,
+  ListNamespacesResponse,
+  CreateNamespaceResponse,
+  DropNamespaceResponse,
+  DescribeNamespaceResponse,
+  RenameTableOptions,
 } from "./connection";
 
 export { Session } from "./native.js";
@@ -73,6 +83,7 @@ export {
   VectorQuery,
   TakeQuery,
   QueryExecutionOptions,
+  ColumnOrdering,
   FullTextSearchOptions,
   RecordBatchIterator,
   FullTextQuery,
@@ -103,6 +114,8 @@ export {
   UpdateOptions,
   OptimizeOptions,
   Version,
+  WriteProgress,
+  LsmWriteSpec,
   ColumnAlteration,
 } from "./table";
 
@@ -117,6 +130,7 @@ export { MergeInsertBuilder, WriteExecutionOptions } from "./merge";
 
 export * as embedding from "./embedding";
 export { permutationBuilder, PermutationBuilder } from "./permutation";
+export { Scannable, ScannableOptions } from "./scannable";
 export * as rerankers from "./rerankers";
 export {
   SchemaLike,
@@ -293,3 +307,197 @@ export async function connect(
   );
   return new LocalConnection(nativeConn);
 }
+
+/**
+ * Configuration for the built-in directory namespace (`"dir"`).
+ *
+ * The directory namespace stores tables under a single root path (local
+ * filesystem or object storage URI). See
+ * {@link https://docs.lancedb.com/namespaces} for the documented surface;
+ * less-common knobs live under {@link DirNamespaceConfig.extraProperties}.
+ */
+export interface DirNamespaceConfig {
+  /** Root path or URI containing the LanceDB tables. */
+  root: string;
+  /**
+   * Whether to maintain a namespace manifest at the root. Required for
+   * child namespaces. Defaults to true on the impl side.
+   */
+  manifestEnabled?: boolean;
+  /**
+   * Additional raw properties passed verbatim to the namespace
+   * implementation (e.g. `storage.*`, `credential_vendor.*`). Typed
+   * fields above take precedence on key collision.
+   */
+  extraProperties?: Record<string, string>;
+}
+
+/**
+ * Configuration for the built-in REST namespace (`"rest"`).
+ *
+ * The REST namespace talks to a remote catalog server over HTTP. See
+ * {@link https://docs.lancedb.com/namespaces} for the documented surface;
+ * less-common knobs (TLS, metrics) live under
+ * {@link RestNamespaceConfig.extraProperties}.
+ */
+export interface RestNamespaceConfig {
+  /** Catalog endpoint URL. */
+  uri: string;
+  /**
+   * HTTP headers forwarded with each request. Keys are passed through
+   * as-is (e.g. `"x-api-key"`, `"Authorization"`).
+   */
+  headers?: Record<string, string>;
+  /**
+   * Additional raw properties passed verbatim to the namespace
+   * implementation (e.g. `tls.*`, `ops_metrics_enabled`, `delimiter`).
+   * Typed fields above take precedence on key collision.
+   */
+  extraProperties?: Record<string, string>;
+}
+
+function dirConfigToProperties(
+  config: DirNamespaceConfig,
+): Record<string, string> {
+  // Spread the whole input so that unknown keys (e.g. a raw `manifest_enabled`
+  // passed via the dynamic-impl path) flow through instead of being dropped.
+  // Typed transformations layer on top.
+  const { manifestEnabled, extraProperties, ...rest } = config;
+  const properties: Record<string, string> = {
+    ...(extraProperties ?? {}),
+    ...(rest as Record<string, string>),
+  };
+  if (manifestEnabled !== undefined) {
+    properties.manifest_enabled = String(manifestEnabled);
+  }
+  return properties;
+}
+
+function restConfigToProperties(
+  config: RestNamespaceConfig,
+): Record<string, string> {
+  const { headers, extraProperties, ...rest } = config;
+  const properties: Record<string, string> = {
+    ...(extraProperties ?? {}),
+    ...(rest as Record<string, string>),
+  };
+  if (headers) {
+    for (const [name, value] of Object.entries(headers)) {
+      properties[`headers.${name}`] = value;
+    }
+  }
+  return properties;
+}
+
+/**
+ * Connect to a LanceDB database through a namespace.
+ *
+ * Unlike {@link connect}, which routes by URI scheme (local path vs.
+ * `db://` cloud), `connectNamespace` always returns a namespace-backed
+ * connection. The `implName` selects the namespace implementation:
+ *
+ * - `"dir"` — directory namespace, configured with {@link DirNamespaceConfig}.
+ * - `"rest"` — remote REST catalog, configured with {@link RestNamespaceConfig}.
+ * - Any other string — full module path for a custom implementation,
+ *   configured with a free-form string-keyed `properties` map.
+ *
+ * @example Typed dir namespace
+ * ```ts
+ * const db = await connectNamespace("dir", { root: "/path/to/db" });
+ * await db.createTable("users", [{ id: 1 }]);
+ * ```
+ *
+ * @example Typed REST namespace with auth headers
+ * ```ts
+ * const db = await connectNamespace("rest", {
+ *   uri: "https://catalog.example.com",
+ *   headers: { "x-api-key": process.env.CATALOG_KEY ?? "" },
+ * });
+ * ```
+ *
+ * @example Custom implementation with raw properties
+ * ```ts
+ * const db = await connectNamespace("my.custom.Namespace", {
+ *   endpoint: "...",
+ * });
+ * ```
+ */
+export function connectNamespace(
+  implName: "dir",
+  config: DirNamespaceConfig,
+  options?: Partial<ConnectNamespaceOptions>,
+): Promise<Connection>;
+/**
+ * Connect through the built-in REST namespace.
+ *
+ * Configured with {@link RestNamespaceConfig}. See the function-level
+ * documentation above for the full surface, examples, and how this
+ * relates to {@link connect}.
+ *
+ * @example
+ * ```ts
+ * const db = await connectNamespace("rest", {
+ *   uri: "https://catalog.example.com",
+ *   headers: { "x-api-key": process.env.CATALOG_KEY ?? "" },
+ * });
+ * ```
+ */
+export function connectNamespace(
+  implName: "rest",
+  config: RestNamespaceConfig,
+  options?: Partial<ConnectNamespaceOptions>,
+): Promise<Connection>;
+/**
+ * Connect through a custom namespace implementation by full module path,
+ * configured with a free-form string-keyed `properties` map. Use the
+ * typed overloads above for the built-in `"dir"` and `"rest"` impls.
+ *
+ * See the function-level documentation above for examples and how this
+ * relates to {@link connect}.
+ *
+ * @example
+ * ```ts
+ * const db = await connectNamespace("my.custom.Namespace", {
+ *   endpoint: "...",
+ * });
+ * ```
+ */
+export function connectNamespace(
+  implName: string,
+  properties: Record<string, string>,
+  options?: Partial<ConnectNamespaceOptions>,
+): Promise<Connection>;
+export async function connectNamespace(
+  implName: string,
+  configOrProperties:
+    | DirNamespaceConfig
+    | RestNamespaceConfig
+    | Record<string, string>,
+  options?: Partial<ConnectNamespaceOptions>,
+): Promise<Connection> {
+  let properties: Record<string, string>;
+  if (implName === "dir") {
+    properties = dirConfigToProperties(
+      configOrProperties as DirNamespaceConfig,
+    );
+  } else if (implName === "rest") {
+    properties = restConfigToProperties(
+      configOrProperties as RestNamespaceConfig,
+    );
+  } else {
+    properties = configOrProperties as Record<string, string>;
+  }
+
+  const finalOptions: ConnectNamespaceOptions = (options ??
+    {}) as ConnectNamespaceOptions;
+  finalOptions.storageOptions = cleanseStorageOptions(
+    finalOptions.storageOptions,
+  );
+
+  const nativeConn = await LanceDbConnection.newWithNamespace(
+    implName,
+    properties,
+    finalOptions,
+  );
+  return new LocalConnection(nativeConn);
+}

nodejs/lancedb/query.ts

@@ -79,6 +79,12 @@ export interface QueryExecutionOptions {
   timeoutMs?: number;
 }
 
+export interface ColumnOrdering {
+  columnName: string;
+  ascending?: boolean;
+  nullsFirst?: boolean;
+}
+
 /**
  * Options that control the behavior of a full text search
  */
@@ -417,6 +423,21 @@ export class StandardQueryBase<
     return this;
   }
 
+  /**
+   * Sort the results by the specified column(s).
+   * @returns This query builder.
+   */
+  orderBy(ordering: ColumnOrdering | ColumnOrdering[]): this {
+    const orderings = Array.isArray(ordering) ? ordering : [ordering];
+    const normalized = orderings.map((o) => ({
+      columnName: o.columnName,
+      ascending: o.ascending ?? true,
+      nullsFirst: o.nullsFirst ?? false,
+    }));
+    this.doCall((inner) => inner.orderBy(normalized));
+    return this;
+  }
+
   /**
    * Skip searching un-indexed data. This can make search faster, but will miss
    * any data that is not yet indexed.

nodejs/lancedb/scannable.ts

@@ -0,0 +1,274 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+import {
+  Table as ArrowTable,
+  RecordBatch,
+  RecordBatchReader,
+  Schema,
+} from "apache-arrow";
+import {
+  fromRecordBatchToStreamBuffer,
+  fromTableToBuffer,
+  makeEmptyTable,
+} from "./arrow";
+import { NapiScannable } from "./native.js";
+
+export interface ScannableOptions {
+  /** Hint about the number of rows. Not validated against the stream. */
+  numRows?: number;
+  /**
+   * Whether the source can be scanned more than once. Defaults to `true` for
+   * `fromTable` / `fromFactory` and `false` for `fromIterable` /
+   * `fromRecordBatchReader`.
+   */
+  rescannable?: boolean;
+}
+
+/**
+ * A data source that can be scanned as a stream of Arrow `RecordBatch`es.
+ *
+ * `Scannable` wraps the schema + optional row count + rescannable flag and
+ * a callback that yields batches one at a time. It is passed to consumers
+ * (e.g. `Table.add`, `createTable`, `mergeInsert` — follow-up work) that
+ * need to pull data without materializing the full dataset in JS memory.
+ *
+ * Batches cross the JS↔Rust boundary as Arrow IPC Stream messages; a fresh
+ * writer serializes each batch, and the Rust side decodes it with
+ * `arrow_ipc::reader::StreamReader`. One batch is in flight at a time.
+ */
+export class Scannable {
+  readonly schema: Schema;
+  readonly numRows: number | null;
+  readonly rescannable: boolean;
+
+  /** @hidden */
+  private readonly native: NapiScannable;
+
+  private constructor(
+    native: NapiScannable,
+    schema: Schema,
+    numRows: number | null,
+    rescannable: boolean,
+  ) {
+    this.native = native;
+    this.schema = schema;
+    this.numRows = numRows;
+    this.rescannable = rescannable;
+  }
+
+  /** @hidden Access the native handle for passing through to Rust consumers. */
+  get inner(): NapiScannable {
+    return this.native;
+  }
+
+  /**
+   * Build a Scannable from an explicit schema and a factory that returns a
+   * fresh batch iterator on each call.
+   *
+   * The factory is invoked once per scan. Each iterator yields
+   * `RecordBatch`es matching the declared schema. Use this when you need
+   * direct control over the pull loop — for example, to wrap a streaming
+   * source whose batches are produced lazily.
+   *
+   * @param schema - The Arrow schema of the produced batches.
+   * @param factory - Called at the start of each scan to produce a batch
+   *   iterator. Must be idempotent when `rescannable` is true.
+   * @param opts - Optional hints. `rescannable` defaults to `true`; set to
+   *   `false` if calling `factory()` twice would not reproduce the same data.
+   */
+  static async fromFactory(
+    schema: Schema,
+    factory: () =>
+      | AsyncIterable<RecordBatch>
+      | Iterable<RecordBatch>
+      | AsyncIterator<RecordBatch>
+      | Iterator<RecordBatch>,
+    opts: ScannableOptions = {},
+  ): Promise<Scannable> {
+    const numRows = opts.numRows ?? null;
+    if (numRows != null && !Number.isInteger(numRows)) {
+      throw new TypeError("numRows must be an integer");
+    }
+    const rescannable = opts.rescannable ?? true;
+
+    let iter: AsyncIterator<RecordBatch> | Iterator<RecordBatch> | null = null;
+    const getNextBatch = async (isStart: boolean): Promise<Buffer | null> => {
+      // `isStart` is true on the first pull of every new scan_as_stream.
+      // Drop any cached iterator so factory() is re-invoked for the next scan
+      if (isStart) {
+        iter = null;
+      }
+      if (iter === null) {
+        iter = normalizeIterator(factory());
+      }
+      const result = await iter.next();
+      if (result.done) {
+        iter = null;
+        return null;
+      }
+      return fromRecordBatchToStreamBuffer(result.value);
+    };
+
+    const schemaBuf = await fromTableToBuffer(makeEmptyTable(schema));
+    const native = new NapiScannable(
+      schemaBuf,
+      numRows,
+      rescannable,
+      getNextBatch,
+    );
+    return new Scannable(native, schema, numRows, rescannable);
+  }
+
+  /**
+   * Build a Scannable from an in-memory Arrow `Table`. Always rescannable;
+   * the table's batches are replayed on each scan.
+   *
+   * The table's row count is authoritative: `opts.numRows` must either be
+   * omitted or equal to `table.numRows`. `opts.rescannable` of `false` is
+   * rejected because in-memory Tables are always rescannable.
+   */
+  static async fromTable(
+    table: ArrowTable,
+    opts: ScannableOptions = {},
+  ): Promise<Scannable> {
+    if (opts.numRows != null && opts.numRows !== table.numRows) {
+      throw new TypeError(
+        `opts.numRows (${opts.numRows}) does not match table.numRows (${table.numRows}). ` +
+          `The table's row count is authoritative; omit numRows or pass the matching value.`,
+      );
+    }
+    if (opts.rescannable === false) {
+      throw new TypeError(
+        `fromTable does not accept rescannable: false. ` +
+          `In-memory Arrow Tables are always rescannable; omit the option or pass true.`,
+      );
+    }
+    return Scannable.fromFactory(table.schema, () => table.batches, {
+      numRows: table.numRows,
+      rescannable: true,
+    });
+  }
+
+  /**
+   * Build a Scannable from an iterable of `RecordBatch`es. `rescannable`
+   * defaults to `false`. Pass an explicit schema so the consumer can
+   * validate before any batch is pulled.
+   *
+   * `opts.rescannable: true` is honest for replayable iterables (Arrays,
+   * Sets, or custom iterables whose `[Symbol.iterator]()` returns a fresh
+   * iterator each call). It is rejected for one-shot iterables (generators,
+   * async generators, or already-an-iterator inputs) because their
+   * `[Symbol.iterator]()` returns the same exhausted object on the second
+   * scan. For replayable sources outside this shape, use
+   * `fromFactory(schema, () => createIter(), { rescannable: true })`.
+   *
+   * Note: when `opts.rescannable` is `true`, the constructor calls
+   * `[Symbol.iterator]()` once on the input to perform the structural check.
+   */
+  static async fromIterable(
+    schema: Schema,
+    iter: AsyncIterable<RecordBatch> | Iterable<RecordBatch>,
+    opts: ScannableOptions = {},
+  ): Promise<Scannable> {
+    if (opts.rescannable === true && isOneShotIterable(iter)) {
+      throw new TypeError(
+        `fromIterable: rescannable: true is not honest for one-shot iterables ` +
+          `(generators, async generators, or iterators where [Symbol.iterator]() ` +
+          `returns the same object). The source would be exhausted after the first scan. ` +
+          `Use fromFactory(schema, () => createIter(), { rescannable: true }) for sources ` +
+          `where each call mints a fresh iterator.`,
+      );
+    }
+    return Scannable.fromFactory(schema, () => iter, {
+      numRows: opts.numRows,
+      rescannable: opts.rescannable ?? false,
+    });
+  }
+
+  /**
+   * Build a Scannable from an Arrow `RecordBatchReader`. A reader can only
+   * be consumed once; `rescannable` defaults to `false`.
+   *
+   * The reader must already be opened (via `.open()`) so its `.schema` is
+   * populated. `RecordBatchReader.from(...)` returns an unopened reader.
+   *
+   * `opts.rescannable: true` is rejected because `RecordBatchReader` is a
+   * self-iterator (its `[Symbol.iterator]()` returns itself), and this
+   * constructor does not call `reader.reset()` between scans, so a second
+   * scan would always see an exhausted reader. For genuinely replayable
+   * sources, use
+   * `fromFactory(schema, () => openReader(), { rescannable: true })`,
+   * which mints a fresh reader on each scan.
+   */
+  static async fromRecordBatchReader(
+    reader: RecordBatchReader,
+    opts: ScannableOptions = {},
+  ): Promise<Scannable> {
+    if (opts.rescannable === true) {
+      throw new TypeError(
+        `fromRecordBatchReader does not accept rescannable: true. ` +
+          `RecordBatchReader is a self-iterator (its [Symbol.iterator]() ` +
+          `returns itself) and would be exhausted after the first scan. ` +
+          `Use fromFactory(schema, () => openReader(), { rescannable: true }) ` +
+          `for sources where each call mints a fresh reader.`,
+      );
+    }
+    return Scannable.fromFactory(reader.schema, () => reader, {
+      numRows: opts.numRows,
+      rescannable: false,
+    });
+  }
+}
+
+function normalizeIterator<T>(
+  source: AsyncIterable<T> | Iterable<T> | AsyncIterator<T> | Iterator<T>,
+): AsyncIterator<T> | Iterator<T> {
+  if (source == null) {
+    throw new TypeError("Scannable factory returned null/undefined");
+  }
+  if (
+    typeof (source as AsyncIterable<T>)[Symbol.asyncIterator] === "function"
+  ) {
+    return (source as AsyncIterable<T>)[Symbol.asyncIterator]();
+  }
+  if (typeof (source as Iterable<T>)[Symbol.iterator] === "function") {
+    return (source as Iterable<T>)[Symbol.iterator]();
+  }
+  // Already an iterator (has `.next`).
+  if (typeof (source as Iterator<T>).next === "function") {
+    return source as Iterator<T>;
+  }
+  throw new TypeError("Scannable factory returned a non-iterable value");
+}
+
+// A "self-iterator" returns the same object from `[Symbol.iterator]()` /
+// `[Symbol.asyncIterator]()`. Generators behave this way, so they exhaust
+// after one pass. Replayable iterables (Array, Set, custom) return a fresh
+// iterator each call. Detection mirrors `normalizeIterator`'s ordering so
+// classification matches scan-time behavior.
+function isOneShotIterable(
+  source: AsyncIterable<unknown> | Iterable<unknown>,
+): boolean {
+  // null/undefined are not one-shot in any meaningful sense; let
+  // `normalizeIterator` raise the actual error at scan time.
+  if (source == null) return false;
+  const ref = source as unknown;
+  if (
+    typeof (source as AsyncIterable<unknown>)[Symbol.asyncIterator] ===
+    "function"
+  ) {
+    const it = (source as AsyncIterable<unknown>)[
+      Symbol.asyncIterator
+    ]() as unknown;
+    return it === ref;
+  }
+  if (typeof (source as Iterable<unknown>)[Symbol.iterator] === "function") {
+    const it = (source as Iterable<unknown>)[Symbol.iterator]() as unknown;
+    return it === ref;
+  }
+  // Already-an-iterator (has `.next` but no `Symbol.iterator`) is by
+  // definition one-shot.
+  if (typeof (source as { next?: unknown }).next === "function") return true;
+  return false;
+}

nodejs/lancedb/table.ts

@@ -46,6 +46,33 @@ import { sanitizeType } from "./sanitize";
 import { IntoSql, toSQL } from "./util";
 export { IndexConfig } from "./native";
 
+/**
+ * Progress snapshot for a write operation, delivered to the `progress`
+ * callback passed to {@link Table.add}.
+ */
+export interface WriteProgress {
+  /** Number of rows written so far. */
+  outputRows: number;
+  /** Number of bytes written so far. */
+  outputBytes: number;
+  /**
+   * Total rows expected, when the input source reports it.
+   *
+   * Always set on the final callback (the one with `done: true`), falling
+   * back to the actual number of rows written when the source could not
+   * report a row count up front.
+   */
+  totalRows?: number;
+  /** Wall-clock seconds since the write started. */
+  elapsedSeconds: number;
+  /** Number of parallel write tasks currently in flight. */
+  activeTasks: number;
+  /** Total number of parallel write tasks (the write parallelism). */
+  totalTasks: number;
+  /** `true` for the final callback; `false` otherwise. */
+  done: boolean;
+}
+
 /**
  * Options for adding data to a table.
  */
@@ -56,6 +83,28 @@ export interface AddDataOptions {
    * If "overwrite" then the new data will replace the existing data in the table.
    */
   mode: "append" | "overwrite";
+
+  /**
+   * Optional callback invoked periodically with write progress.
+   *
+   * The callback is fired once per batch written and once more with
+   * `done: true` when the write completes. Calls are dispatched
+   * asynchronously to the JS event loop and never block the write — a slow
+   * callback will queue events rather than back-pressure the writer.
+   *
+   * Errors thrown from the callback are logged with `console.warn` and
+   * swallowed — they do not abort the write.
+   *
+   * @example
+   * ```ts
+   * await table.add(data, {
+   *   progress: (p) => {
+   *     console.log(`${p.outputRows}/${p.totalRows ?? "?"} rows`);
+   *   },
+   * });
+   * ```
+   */
+  progress: (progress: WriteProgress) => void;
 }
 
 export interface UpdateOptions {
@@ -106,6 +155,27 @@ export interface Version {
   metadata: Record<string, string>;
 }
 
+/**
+ * Specification selecting Lance's MemWAL LSM-style write path for
+ * `mergeInsert`.
+ *
+ * `specType` is `"bucket"`, `"identity"`, or `"unsharded"`. For `"bucket"`,
+ * `column` and `numBuckets` are required; for `"identity"`, `column` is
+ * required.
+ */
+export interface LsmWriteSpec {
+  /** One of `"bucket"`, `"identity"`, or `"unsharded"`. */
+  specType: "bucket" | "identity" | "unsharded";
+  /** Bucket and identity variants: the sharding column. */
+  column?: string;
+  /** Bucket variant: the number of buckets, in `[1, 1024]`. */
+  numBuckets?: number;
+  /** Names of indexes the MemWAL should keep up to date during writes. */
+  maintainedIndexes?: string[];
+  /** Default `ShardWriter` configuration recorded in the MemWAL index. */
+  writerConfigDefaults?: Record<string, string>;
+}
+
 /**
  * A Table is a collection of Records in a LanceDB Database.
  *
@@ -449,6 +519,54 @@ export abstract class Table {
    * containing the new version number of the table after dropping the columns.
    */
   abstract dropColumns(columnNames: string[]): Promise<DropColumnsResult>;
+  /**
+   * Set the unenforced primary key for this table to a single column.
+   *
+   * "Unenforced" means LanceDB does not check uniqueness on writes; the
+   * column is recorded in the schema as the primary key for use by features
+   * such as `merge_insert`. Only single-column primary keys are supported,
+   * and the key cannot be changed once set.
+   * @param {string | string[]} columns The primary key column. A one-element
+   * array is also accepted; passing more than one column is rejected.
+   * @returns {Promise<void>}
+   */
+  abstract setUnenforcedPrimaryKey(columns: string | string[]): Promise<void>;
+  /**
+   * Install an {@link LsmWriteSpec} on this table, selecting Lance's MemWAL
+   * LSM-style write path for future `mergeInsert` calls.
+   *
+   * `LsmWriteSpec` chooses one of three sharding strategies via `specType`:
+   *
+   * - `"bucket"` — hash-bucket writes by the single-column unenforced primary
+   *   key (`column` and `numBuckets` required).
+   * - `"identity"` — shard by the raw value of a scalar `column`.
+   * - `"unsharded"` — route every write to a single shard.
+   *
+   * All variants require the table to have an unenforced primary key
+   * ({@link Table#setUnenforcedPrimaryKey}); bucket sharding additionally
+   * requires it to be the single column being bucketed.
+   * @param {LsmWriteSpec} spec The sharding spec to install.
+   * @returns {Promise<void>}
+   * @example
+   * ```ts
+   * await table.setUnenforcedPrimaryKey("id");
+   * await table.setLsmWriteSpec({
+   *   specType: "bucket",
+   *   column: "id",
+   *   numBuckets: 16,
+   *   maintainedIndexes: ["id_idx"],
+   * });
+   * ```
+   */
+  abstract setLsmWriteSpec(spec: LsmWriteSpec): Promise<void>;
+  /**
+   * Remove the {@link LsmWriteSpec} from this table, reverting to the standard
+   * `mergeInsert` write path.
+   *
+   * Errors if no spec is currently set.
+   * @returns {Promise<void>}
+   */
+  abstract unsetLsmWriteSpec(): Promise<void>;
   /** Retrieve the version of the table */
 
   abstract version(): Promise<number>;
@@ -636,7 +754,20 @@ export class LocalTable extends Table {
     const schema = await this.schema();
 
     const buffer = await fromDataToBuffer(data, undefined, schema);
-    return await this.inner.add(buffer, mode);
+    // Wrap the user callback so a thrown error doesn't surface as an
+    // unhandled exception (the callback fires from a napi threadsafe
+    // function — exceptions there crash the process).
+    const userProgress = options?.progress;
+    const progress = userProgress
+      ? (p: WriteProgress) => {
+          try {
+            userProgress(p);
+          } catch (e) {
+            console.warn("Table.add progress callback threw:", e);
+          }
+        }
+      : undefined;
+    return await this.inner.add(buffer, mode, progress);
   }
 
   async update(
@@ -897,6 +1028,19 @@ export class LocalTable extends Table {
     return await this.inner.dropColumns(columnNames);
   }
 
+  async setUnenforcedPrimaryKey(columns: string | string[]): Promise<void> {
+    const cols = typeof columns === "string" ? [columns] : columns;
+    return await this.inner.setUnenforcedPrimaryKey(cols);
+  }
+
+  async setLsmWriteSpec(spec: LsmWriteSpec): Promise<void> {
+    return await this.inner.setLsmWriteSpec(spec);
+  }
+
+  async unsetLsmWriteSpec(): Promise<void> {
+    return await this.inner.unsetLsmWriteSpec();
+  }
+
   async version(): Promise<number> {
     return await this.inner.version();
   }

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

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

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.28.0-beta.11",
+  "version": "0.29.1-beta.0",
   "os": [
     "win32"
   ],

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.28.0-beta.11",
+	"version": "0.29.1-beta.0",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.28.0-beta.11",
+  "version": "0.29.1-beta.0",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",
@@ -38,15 +38,15 @@
     "url": "https://github.com/lancedb/lancedb"
   },
   "devDependencies": {
-    "@aws-sdk/client-dynamodb": "^3.33.0",
-    "@aws-sdk/client-kms": "^3.33.0",
-    "@aws-sdk/client-s3": "^3.33.0",
+    "@aws-sdk/client-dynamodb": "3.1003.0",
+    "@aws-sdk/client-kms": "3.1003.0",
+    "@aws-sdk/client-s3": "3.1003.0",
     "@biomejs/biome": "^1.7.3",
     "@jest/globals": "^29.7.0",
-    "@napi-rs/cli": "^3.5.1",
+    "@napi-rs/cli": "3.5.1",
     "@types/axios": "^0.14.0",
     "@types/jest": "^29.1.2",
-    "@types/node": "^22.7.4",
+    "@types/node": "22.7.4",
     "@types/tmp": "^0.2.6",
     "apache-arrow-15": "npm:apache-arrow@15.0.0",
     "apache-arrow-16": "npm:apache-arrow@16.0.0",
@@ -57,9 +57,9 @@
     "shx": "^0.3.4",
     "tmp": "^0.2.3",
     "ts-jest": "^29.1.2",
-    "typedoc": "^0.26.4",
-    "typedoc-plugin-markdown": "^4.2.1",
-    "typescript": "^5.5.4",
+    "typedoc": "0.26.4",
+    "typedoc-plugin-markdown": "4.2.1",
+    "typescript": "5.5.4",
     "typescript-eslint": "^7.1.0"
   },
   "ava": {
@@ -68,15 +68,16 @@
   "engines": {
     "node": ">= 18"
   },
+  "packageManager": "pnpm@11.1.1",
   "cpu": ["x64", "arm64"],
   "os": ["darwin", "linux", "win32"],
   "scripts": {
     "artifacts": "napi artifacts",
     "build:debug": "napi build --platform --dts ../lancedb/native.d.ts --js ../lancedb/native.js --output-dir lancedb",
-    "postbuild:debug": "shx mkdir -p dist && shx cp lancedb/*.node dist/",
+    "postbuild:debug": "shx mkdir -p dist && shx cp lancedb/*.node dist/ && node -e \"require('fs').writeFileSync('dist/package.json', JSON.stringify({name:'@lancedb/lancedb',type:'commonjs'}))\"",
     "build:release": "napi build --platform --release --dts ../lancedb/native.d.ts --js ../lancedb/native.js --output-dir dist",
-    "build": "npm run build:debug && npm run tsc",
-    "build-release": "npm run build:release && npm run tsc",
+    "build": "pnpm build:debug && pnpm tsc",
+    "build-release": "pnpm build:release && pnpm tsc",
     "tsc": "tsc -b",
     "posttsc": "shx cp lancedb/native.d.ts dist/native.d.ts",
     "lint-ci": "biome ci .",
@@ -86,7 +87,7 @@
     "lint-fix": "biome check --write . && biome format --write .",
     "prepublishOnly": "napi prepublish -t npm",
     "test": "jest --verbose",
-    "integration": "S3_TEST=1 npm run test",
+    "integration": "S3_TEST=1 pnpm test",
     "universal": "napi universalize",
     "version": "napi version"
   },
@@ -94,8 +95,8 @@
     "reflect-metadata": "^0.2.2"
   },
   "optionalDependencies": {
-    "@huggingface/transformers": "^3.0.2",
-    "openai": "^4.29.2"
+    "@huggingface/transformers": "3.0.2",
+    "openai": "4.29.2"
   },
   "peerDependencies": {
     "apache-arrow": ">=15.0.0 <=18.1.0"

nodejs/pnpm-workspace.yaml

@@ -0,0 +1,18 @@
+# Flat node_modules layout. The @napi-rs/cli build step fails to locate
+# the cdylib artifact under pnpm's isolated layout; the hoisted linker
+# mirrors npm's structure and unblocks the native build.
+nodeLinker: hoisted
+
+# Block resolution of versions less than 24h old (Shai-Hulud window).
+# This is the pnpm 11 default but pinned here so it's visible to
+# reviewers and survives a future pnpm major flipping the default.
+minimumReleaseAge: 1440
+
+# Fail install if a transitive dep tries to run an unapproved script.
+strictDepBuilds: true
+
+allowBuilds:
+  '@biomejs/biome': true
+  onnxruntime-node: true
+  protobufjs: true
+  sharp: true

nodejs/src/connection.rs

@@ -8,12 +8,16 @@ use lancedb::database::{CreateTableMode, Database};
 use napi::bindgen_prelude::*;
 use napi_derive::*;
 
+use crate::ConnectNamespaceOptions;
 use crate::ConnectionOptions;
 use crate::error::NapiErrorExt;
 use crate::header::JsHeaderProvider;
 use crate::table::Table;
-use lancedb::connection::{ConnectBuilder, Connection as LanceDBConnection};
+use lancedb::connection::{ConnectBuilder, Connection as LanceDBConnection, connect_namespace};
 
+use lance_namespace::models::{
+    CreateNamespaceRequest, DescribeNamespaceRequest, DropNamespaceRequest, ListNamespacesRequest,
+};
 use lancedb::ipc::{ipc_file_to_batches, ipc_file_to_schema};
 
 #[napi]
@@ -21,6 +25,29 @@ pub struct Connection {
     inner: Option<LanceDBConnection>,
 }
 
+#[napi(object)]
+pub struct DescribeNamespaceResponse {
+    pub properties: Option<HashMap<String, String>>,
+}
+
+#[napi(object)]
+pub struct ListNamespacesResponse {
+    pub namespaces: Vec<String>,
+    pub page_token: Option<String>,
+}
+
+#[napi(object)]
+pub struct CreateNamespaceResponse {
+    pub properties: Option<HashMap<String, String>>,
+    pub transaction_id: Option<String>,
+}
+
+#[napi(object)]
+pub struct DropNamespaceResponse {
+    pub properties: Option<HashMap<String, String>>,
+    pub transaction_id: Option<Vec<String>>,
+}
+
 impl Connection {
     pub(crate) fn inner_new(inner: LanceDBConnection) -> Self {
         Self { inner: Some(inner) }
@@ -106,6 +133,39 @@ impl Connection {
         Ok(Self::inner_new(builder.execute().await.default_error()?))
     }
 
+    /// Create a new Connection instance backed by a namespace implementation.
+    #[napi(factory)]
+    pub async fn new_with_namespace(
+        impl_name: String,
+        properties: HashMap<String, String>,
+        options: ConnectNamespaceOptions,
+    ) -> napi::Result<Self> {
+        if impl_name.is_empty() {
+            return Err(napi::Error::from_reason(
+                "implName must be a non-empty string",
+            ));
+        }
+
+        let mut builder = connect_namespace(&impl_name, properties);
+        if let Some(interval) = options.read_consistency_interval {
+            builder =
+                builder.read_consistency_interval(std::time::Duration::from_secs_f64(interval));
+        }
+        if let Some(storage_options) = options.storage_options {
+            for (key, value) in storage_options {
+                builder = builder.storage_option(key, value);
+            }
+        }
+        if let Some(namespace_client_properties) = options.namespace_client_properties {
+            builder = builder.namespace_client_properties(namespace_client_properties);
+        }
+        if let Some(session) = options.session {
+            builder = builder.session(session.inner.clone());
+        }
+
+        Ok(Self::inner_new(builder.execute().await.default_error()?))
+    }
+
     #[napi]
     pub fn display(&self) -> napi::Result<String> {
         Ok(self.get_inner()?.to_string())
@@ -273,4 +333,149 @@ impl Connection {
         let ns = namespace_path.unwrap_or_default();
         self.get_inner()?.drop_all_tables(&ns).await.default_error()
     }
+
+    #[napi(catch_unwind)]
+    /// Describe a namespace and return its properties.
+    pub async fn describe_namespace(
+        &self,
+        namespace_path: Vec<String>,
+    ) -> napi::Result<DescribeNamespaceResponse> {
+        let req = DescribeNamespaceRequest {
+            id: Some(namespace_path),
+            ..Default::default()
+        };
+        let resp = self
+            .get_inner()?
+            .describe_namespace(req)
+            .await
+            .default_error()?;
+        Ok(DescribeNamespaceResponse {
+            properties: resp.properties,
+        })
+    }
+
+    #[napi(catch_unwind)]
+    /// List child namespaces under the given namespace path
+    pub async fn list_namespaces(
+        &self,
+        namespace_path: Option<Vec<String>>,
+        page_token: Option<String>,
+        limit: Option<u32>,
+    ) -> napi::Result<ListNamespacesResponse> {
+        let req = ListNamespacesRequest {
+            id: namespace_path,
+            page_token,
+            limit: limit.map(|l| l as i32),
+            ..Default::default()
+        };
+        let resp = self
+            .get_inner()?
+            .list_namespaces(req)
+            .await
+            .default_error()?;
+        Ok(ListNamespacesResponse {
+            namespaces: resp.namespaces,
+            page_token: resp.page_token,
+        })
+    }
+
+    #[napi(catch_unwind)]
+    /// Create a new namespace with optional properties.
+    pub async fn create_namespace(
+        &self,
+        namespace_path: Vec<String>,
+        mode: Option<String>,
+        properties: Option<HashMap<String, String>>,
+    ) -> napi::Result<CreateNamespaceResponse> {
+        let mode_str = mode
+            .map(|m| match m.to_lowercase().as_str() {
+                "create" => Ok("Create".to_string()),
+                "exist_ok" => Ok("ExistOk".to_string()),
+                "overwrite" => Ok("Overwrite".to_string()),
+                _ => Err(napi::Error::from_reason(format!(
+                    "Invalid mode '{}': expected one of 'create', 'exist_ok', 'overwrite'",
+                    m
+                ))),
+            })
+            .transpose()?;
+        let req = CreateNamespaceRequest {
+            id: Some(namespace_path),
+            mode: mode_str,
+            properties,
+            ..Default::default()
+        };
+        let resp = self
+            .get_inner()?
+            .create_namespace(req)
+            .await
+            .default_error()?;
+        Ok(CreateNamespaceResponse {
+            properties: resp.properties,
+            transaction_id: resp.transaction_id,
+        })
+    }
+
+    #[napi(catch_unwind)]
+    /// Drop a namespace.
+    pub async fn drop_namespace(
+        &self,
+        namespace_path: Vec<String>,
+        mode: Option<String>,
+        behavior: Option<String>,
+    ) -> napi::Result<DropNamespaceResponse> {
+        let mode_str = mode
+            .map(|m| match m.to_lowercase().as_str() {
+                "skip" => Ok("Skip".to_string()),
+                "fail" => Ok("Fail".to_string()),
+                _ => Err(napi::Error::from_reason(format!(
+                    "Invalid mode '{}': expected one of 'skip', 'fail'",
+                    m
+                ))),
+            })
+            .transpose()?;
+        let behavior_str = behavior
+            .map(|b| match b.to_lowercase().as_str() {
+                "restrict" => Ok("Restrict".to_string()),
+                "cascade" => Ok("Cascade".to_string()),
+                _ => Err(napi::Error::from_reason(format!(
+                    "Invalid behavior '{}': expected one of 'restrict', 'cascade'",
+                    b
+                ))),
+            })
+            .transpose()?;
+        let req = DropNamespaceRequest {
+            id: Some(namespace_path),
+            mode: mode_str,
+            behavior: behavior_str,
+            ..Default::default()
+        };
+        let resp = self
+            .get_inner()?
+            .drop_namespace(req)
+            .await
+            .default_error()?;
+        Ok(DropNamespaceResponse {
+            properties: resp.properties,
+            transaction_id: resp.transaction_id,
+        })
+    }
+
+    /// Rename a table. `current_namespace_path` and `new_namespace_path` default to
+    /// the root namespace when omitted; the caller is expected to either pass both
+    /// or pass neither.
+    #[napi(catch_unwind)]
+    pub async fn rename_table(
+        &self,
+        current_name: String,
+        new_name: String,
+        current_namespace_path: Option<Vec<String>>,
+        new_namespace_path: Option<Vec<String>>,
+    ) -> napi::Result<()> {
+        let cur_ns = current_namespace_path.unwrap_or_default();
+        let new_ns = new_namespace_path.unwrap_or_default();
+        self.get_inner()?
+            .rename_table(&current_name, &new_name, &cur_ns, &new_ns)
+            .await
+            .default_error()
+    }
 }

nodejs/src/lib.rs

@@ -16,6 +16,7 @@ pub mod permutation;
 mod query;
 pub mod remote;
 mod rerankers;
+mod scannable;
 mod session;
 mod table;
 mod util;
@@ -67,6 +68,26 @@ pub struct OpenTableOptions {
     pub storage_options: Option<HashMap<String, String>>,
 }
 
+#[napi(object)]
+#[derive(Debug)]
+pub struct ConnectNamespaceOptions {
+    /// The interval, in seconds, at which to check for updates to the table
+    /// from other processes. If None, then consistency is not checked. For
+    /// performance reasons, this is the default. For strong consistency, set
+    /// this to zero seconds. Then every read will check for updates from other
+    /// processes. As a compromise, you can set this to a non-zero value for
+    /// eventual consistency.
+    pub read_consistency_interval: Option<f64>,
+    /// Configuration for object storage. The available options are described
+    /// at https://docs.lancedb.com/storage/
+    pub storage_options: Option<HashMap<String, String>>,
+    /// Extra properties for the backing namespace client.
+    pub namespace_client_properties: Option<HashMap<String, String>>,
+    /// The session to use for this connection. Holds shared caches and other
+    /// session-specific state.
+    pub session: Option<session::Session>,
+}
+
 #[napi_derive::module_init]
 fn init() {
     let env = Env::new()

nodejs/src/query.rs

@@ -3,6 +3,12 @@
 
 use std::sync::Arc;
 
+use crate::error::NapiErrorExt;
+use crate::error::convert_error;
+use crate::iterator::RecordBatchIterator;
+use crate::rerankers::RerankHybridCallbackArgs;
+use crate::rerankers::Reranker;
+use crate::util::{parse_distance_type, schema_to_buffer};
 use arrow_array::{
     Array, Float16Array as ArrowFloat16Array, Float32Array as ArrowFloat32Array,
     Float64Array as ArrowFloat64Array, UInt8Array as ArrowUInt8Array,
@@ -19,16 +25,27 @@ use lancedb::query::QueryBase;
 use lancedb::query::QueryExecutionOptions;
 use lancedb::query::Select;
 use lancedb::query::TakeQuery as LanceDbTakeQuery;
-use lancedb::query::VectorQuery as LanceDbVectorQuery;
+use lancedb::query::{ColumnOrdering as LanceDbColumnOrdering, VectorQuery as LanceDbVectorQuery};
 use napi::bindgen_prelude::*;
 use napi_derive::napi;
 
-use crate::error::NapiErrorExt;
-use crate::error::convert_error;
-use crate::iterator::RecordBatchIterator;
-use crate::rerankers::RerankHybridCallbackArgs;
-use crate::rerankers::Reranker;
-use crate::util::{parse_distance_type, schema_to_buffer};
+#[napi(object)]
+pub struct ColumnOrdering {
+    pub ascending: bool,
+    pub nulls_first: bool,
+    pub column_name: String,
+}
+
+impl From<ColumnOrdering> for LanceDbColumnOrdering {
+    fn from(value: ColumnOrdering) -> Self {
+        match (value.ascending, value.nulls_first) {
+            (true, true) => Self::asc_nulls_first(value.column_name),
+            (true, false) => Self::asc_nulls_last(value.column_name),
+            (false, true) => Self::desc_nulls_first(value.column_name),
+            (false, false) => Self::desc_nulls_last(value.column_name),
+        }
+    }
+}
 
 fn bytes_to_arrow_array(data: Uint8Array, dtype: String) -> napi::Result<Arc<dyn Array>> {
     let buf = arrow_buffer::Buffer::from(data.to_vec());
@@ -128,6 +145,18 @@ impl Query {
         self.inner = self.inner.clone().with_row_id();
     }
 
+    #[napi]
+    pub fn order_by(&mut self, ordering: Option<Vec<ColumnOrdering>>) -> napi::Result<()> {
+        let ordering = ordering.map(|ordering| {
+            ordering
+                .into_iter()
+                .map(LanceDbColumnOrdering::from)
+                .collect()
+        });
+        self.inner = self.inner.clone().order_by(ordering);
+        Ok(())
+    }
+
     #[napi(catch_unwind)]
     pub async fn output_schema(&self) -> napi::Result<Buffer> {
         let schema = self.inner.output_schema().await.default_error()?;
@@ -328,6 +357,18 @@ impl VectorQuery {
         Ok(())
     }
 
+    #[napi]
+    pub fn order_by(&mut self, ordering: Option<Vec<ColumnOrdering>>) -> napi::Result<()> {
+        let ordering = ordering.map(|ordering| {
+            ordering
+                .into_iter()
+                .map(LanceDbColumnOrdering::from)
+                .collect()
+        });
+        self.inner = self.inner.clone().order_by(ordering);
+        Ok(())
+    }
+
     #[napi(catch_unwind)]
     pub async fn output_schema(&self) -> napi::Result<Buffer> {
         let schema = self.inner.output_schema().await.default_error()?;

nodejs/src/scannable.rs

@@ -0,0 +1,253 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! NodeJS binding for the [`lancedb::data::scannable::Scannable`] trait.
+//!
+//! The JS side supplies a `getNextBatch(isStart)` callback that returns the
+//! next Arrow `RecordBatch` encoded as a self-contained Arrow IPC Stream
+//! message (schema message + record batch message + EOS marker) wrapped in a
+//! `Buffer`, or `null` when the stream is exhausted. The Rust side parses
+//! each buffer with `arrow_ipc::reader::StreamReader`, validates every
+//! standalone batch stream against the declared schema, and yields decoded
+//! `RecordBatch`es as a [`SendableRecordBatchStream`].
+//!
+//! `isStart` is `true` on the first `getNextBatch` call of each new
+//! `scan_as_stream` and `false` thereafter. JS uses it to drop any cached
+//! iterator and re-invoke its factory at scan boundaries, so retries
+//! triggered by mid-stream failures restart at batch 0.
+
+use std::io::Cursor;
+use std::sync::Arc;
+
+use arrow_array::RecordBatch;
+use arrow_ipc::reader::StreamReader;
+use arrow_schema::SchemaRef;
+use futures::stream::once;
+use lancedb::arrow::{SendableRecordBatchStream, SimpleRecordBatchStream};
+use lancedb::data::scannable::Scannable as LanceScannable;
+use lancedb::ipc::ipc_file_to_schema;
+use lancedb::{Error, Result as LanceResult};
+use napi::bindgen_prelude::*;
+use napi::threadsafe_function::ThreadsafeFunction;
+use napi_derive::napi;
+
+/// Threadsafe handle to the JS `getNextBatch` callback. The callback takes a
+/// single boolean `isStart` (`true` on the first call of each new scan) and
+/// returns a Promise that resolves to a `Buffer` containing one IPC Stream
+/// message, or `null` at end-of-stream.
+type GetNextBatchFn = ThreadsafeFunction<bool, Promise<Option<Buffer>>, bool, Status, false>;
+
+/// A Rust-side view of a JS-constructed `Scannable`.
+///
+/// Held in JS as the return value of the `Scannable` class constructor. When
+/// passed to a consumer that accepts `impl lancedb::data::scannable::Scannable`,
+/// the consumer invokes `scan_as_stream()` to pull batches through the JS
+/// callback.
+#[napi]
+pub struct NapiScannable {
+    schema: SchemaRef,
+    num_rows: Option<usize>,
+    rescannable: bool,
+    // `ThreadsafeFunction` is not `Clone`; wrap in `Arc` so the stream
+    // returned by `scan_as_stream` can own a handle independent of `self`.
+    get_next_batch: Arc<GetNextBatchFn>,
+    // Tracks whether a scan has already started; used to enforce one-shot
+    // semantics on non-rescannable sources.
+    scanned: bool,
+}
+
+#[napi]
+impl NapiScannable {
+    /// Construct a new `NapiScannable`.
+    ///
+    /// - `schema_buf` — Arrow IPC File buffer carrying only the schema (no batches).
+    /// - `num_rows` — optional row count hint; not validated against the stream.
+    /// - `rescannable` — whether `get_next_batch` may be re-driven after the
+    ///   scan completes.
+    /// - `get_next_batch` -- JS callback that yields the next batch as an Arrow
+    ///   IPC Stream message wrapped in a `Buffer`, or `null` at EOF. The
+    ///   `isStart` argument is `true` on the first call of each new scan;
+    ///   JS uses it to discard any cached iterator before pulling.
+    #[napi(constructor)]
+    pub fn new(
+        schema_buf: Buffer,
+        num_rows: Option<i64>,
+        rescannable: bool,
+        get_next_batch: Function<bool, Promise<Option<Buffer>>>,
+    ) -> napi::Result<Self> {
+        let schema = ipc_file_to_schema(schema_buf.to_vec())
+            .map_err(|e| napi::Error::from_reason(format!("Invalid schema buffer: {}", e)))?;
+        let num_rows = num_rows
+            .map(|n| {
+                usize::try_from(n)
+                    .map_err(|_| napi::Error::from_reason("num_rows must be non-negative"))
+            })
+            .transpose()?;
+        let get_next_batch = Arc::new(get_next_batch.build_threadsafe_function().build()?);
+        Ok(Self {
+            schema,
+            num_rows,
+            rescannable,
+            get_next_batch,
+            scanned: false,
+        })
+    }
+}
+
+impl std::fmt::Debug for NapiScannable {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("NapiScannable")
+            .field("schema", &self.schema)
+            .field("num_rows", &self.num_rows)
+            .field("rescannable", &self.rescannable)
+            .finish()
+    }
+}
+
+impl LanceScannable for NapiScannable {
+    fn schema(&self) -> SchemaRef {
+        self.schema.clone()
+    }
+
+    fn scan_as_stream(&mut self) -> SendableRecordBatchStream {
+        let schema = self.schema.clone();
+
+        // One-shot enforcement for non-rescannable sources: return a stream
+        // whose first item is an error.
+        if self.scanned && !self.rescannable {
+            let err_stream = once(async {
+                Err(Error::InvalidInput {
+                    message: "Scannable has already been consumed (non-rescannable source)"
+                        .to_string(),
+                })
+            });
+            return Box::pin(SimpleRecordBatchStream::new(err_stream, schema));
+        }
+        self.scanned = true;
+
+        let tsfn = Arc::clone(&self.get_next_batch);
+        let declared_schema = schema.clone();
+
+        // State threaded through the unfold. `is_first_pull` starts true so
+        // the first call into JS signals a new-scan boundary; JS uses it to
+        // reset any cached iterator before factory()-ing a fresh one.
+        let initial = State {
+            tsfn,
+            batch_index: 0,
+            declared_schema,
+            errored: false,
+            is_first_pull: true,
+        };
+
+        let stream = futures::stream::unfold(initial, |mut state| async move {
+            if state.errored {
+                return None;
+            }
+
+            // Pull the next IPC Stream buffer from JS. `is_first_pull` is
+            // consumed here and cleared so subsequent pulls continue the
+            // same scan rather than restarting it.
+            let is_start = state.is_first_pull;
+            state.is_first_pull = false;
+            let buf = match pull_next(&state.tsfn, is_start).await {
+                Ok(Some(buf)) => buf,
+                Ok(None) => return None,
+                Err(e) => {
+                    state.errored = true;
+                    return Some((Err(e), state));
+                }
+            };
+
+            match decode_one_batch(buf.as_ref(), &state.declared_schema) {
+                Ok(batch) => {
+                    state.batch_index += 1;
+                    Some((Ok(batch), state))
+                }
+                Err(e) => {
+                    let tagged = Error::Runtime {
+                        message: format!(
+                            "[scannable/rust-bridge] failure at batch index {}: {}",
+                            state.batch_index, e
+                        ),
+                    };
+                    state.errored = true;
+                    Some((Err(tagged), state))
+                }
+            }
+        });
+
+        Box::pin(SimpleRecordBatchStream::new(stream, schema))
+    }
+
+    fn num_rows(&self) -> Option<usize> {
+        self.num_rows
+    }
+
+    fn rescannable(&self) -> bool {
+        self.rescannable
+    }
+}
+
+struct State {
+    tsfn: Arc<GetNextBatchFn>,
+    batch_index: usize,
+    declared_schema: SchemaRef,
+    errored: bool,
+    /// True for the very first pull of a new scan. Forwarded to JS so the
+    /// callback can drop any cached iterator and call its factory fresh,
+    /// which makes rescannable sources restart at batch 0 even when the
+    /// previous scan ended mid-stream.
+    is_first_pull: bool,
+}
+
+/// Invoke the JS callback and await its Promise. `is_start` is forwarded to
+/// the JS side as the `isStart` argument so it can reset its iterator at the
+/// scan boundary. Errors on the JS side surface here as rejected promises
+/// and are tunneled back as `lancedb::Error::Runtime`.
+async fn pull_next(tsfn: &GetNextBatchFn, is_start: bool) -> LanceResult<Option<Buffer>> {
+    let promise = tsfn
+        .call_async(is_start)
+        .await
+        .map_err(|e| Error::Runtime {
+            message: format!(
+                "[scannable/js-factory] napi error status={}, reason={}",
+                e.status, e.reason
+            ),
+        })?;
+    promise.await.map_err(|e| Error::Runtime {
+        message: format!(
+            "[scannable/js-iterator] napi error status={}, reason={}",
+            e.status, e.reason
+        ),
+    })
+}
+
+/// Decode one IPC Stream buffer (schema + batch + EOS) into a `RecordBatch`.
+/// Each buffer is a standalone IPC stream, so every decoded stream schema must
+/// match the one declared at construction.
+fn decode_one_batch(buf: &[u8], declared: &SchemaRef) -> LanceResult<RecordBatch> {
+    let reader = StreamReader::try_new(Cursor::new(buf), None).map_err(|e| Error::Runtime {
+        message: format!("failed to open IPC stream reader: {}", e),
+    })?;
+
+    let actual = reader.schema();
+    if actual.as_ref() != declared.as_ref() {
+        return Err(Error::InvalidInput {
+            message: format!(
+                "declared schema does not match stream schema: declared={:?} actual={:?}",
+                declared, actual
+            ),
+        });
+    }
+
+    let mut iter = reader;
+    let batch = iter
+        .next()
+        .ok_or_else(|| Error::Runtime {
+            message: "IPC stream contained schema but no record batch".to_string(),
+        })?
+        .map_err(|e| Error::Runtime {
+            message: format!("failed to decode record batch: {}", e),
+        })?;
+    Ok(batch)
+}

nodejs/src/table.rs

@@ -9,6 +9,7 @@ use lancedb::table::{
     OptimizeAction, OptimizeOptions, Table as LanceDbTable,
 };
 use napi::bindgen_prelude::*;
+use napi::threadsafe_function::{ThreadsafeFunction, ThreadsafeFunctionCallMode};
 use napi_derive::napi;
 
 use crate::error::NapiErrorExt;
@@ -67,8 +68,16 @@ impl Table {
         schema_to_buffer(&schema)
     }
 
-    #[napi(catch_unwind)]
-    pub async fn add(&self, buf: Buffer, mode: String) -> napi::Result<AddResult> {
+    #[napi(
+        catch_unwind,
+        ts_args_type = "buf: Buffer, mode: string, progressCallback?: (progress: WriteProgressInfo) => void"
+    )]
+    pub async fn add(
+        &self,
+        buf: Buffer,
+        mode: String,
+        progress_callback: Option<ProgressFn>,
+    ) -> napi::Result<AddResult> {
         let batches = ipc_file_to_batches(buf.to_vec())
             .map_err(|e| napi::Error::from_reason(format!("Failed to read IPC file: {}", e)))?;
         let batches = batches
@@ -92,6 +101,19 @@ impl Table {
             return Err(napi::Error::from_reason(format!("Invalid mode: {}", mode)));
         };
 
+        if let Some(tsfn) = progress_callback {
+            op = op.progress(move |p| {
+                // NonBlocking: dispatch onto the JS event loop without
+                // blocking the writer thread.  With napi-rs's default
+                // unbounded queue, events are not dropped — a slow JS
+                // callback will just queue them.
+                tsfn.call(
+                    WriteProgressInfo::from(p),
+                    ThreadsafeFunctionCallMode::NonBlocking,
+                );
+            });
+        }
+
         let res = op.execute().await.default_error()?;
         Ok(res.into())
     }
@@ -344,6 +366,31 @@ impl Table {
         Ok(res.into())
     }
 
+    #[napi(catch_unwind)]
+    pub async fn set_unenforced_primary_key(&self, columns: Vec<String>) -> napi::Result<()> {
+        self.inner_ref()?
+            .set_unenforced_primary_key(columns)
+            .await
+            .default_error()
+    }
+
+    #[napi(catch_unwind)]
+    pub async fn set_lsm_write_spec(&self, spec: LsmWriteSpec) -> napi::Result<()> {
+        let native_spec = lancedb::table::LsmWriteSpec::try_from(spec)?;
+        self.inner_ref()?
+            .set_lsm_write_spec(native_spec)
+            .await
+            .default_error()
+    }
+
+    #[napi(catch_unwind)]
+    pub async fn unset_lsm_write_spec(&self) -> napi::Result<()> {
+        self.inner_ref()?
+            .unset_lsm_write_spec()
+            .await
+            .default_error()
+    }
+
     #[napi(catch_unwind)]
     pub async fn version(&self) -> napi::Result<i64> {
         self.inner_ref()?
@@ -538,6 +585,63 @@ impl From<lancedb::index::IndexConfig> for IndexConfig {
     }
 }
 
+/// Specification selecting Lance's MemWAL LSM-style write path for
+/// `mergeInsert`.
+///
+/// `specType` must be `"bucket"`, `"identity"`, or `"unsharded"`. For
+/// `"bucket"`, `column` and `numBuckets` are required; for `"identity"`,
+/// `column` is required.
+#[napi(object)]
+#[derive(Clone, Debug)]
+pub struct LsmWriteSpec {
+    /// One of `"bucket"`, `"identity"`, or `"unsharded"`.
+    pub spec_type: String,
+    /// Bucket and identity variants: the sharding column.
+    pub column: Option<String>,
+    /// Bucket variant: the number of buckets, in `[1, 1024]`.
+    pub num_buckets: Option<u32>,
+    /// Names of indexes the MemWAL should keep up to date during writes.
+    pub maintained_indexes: Option<Vec<String>>,
+    /// Default `ShardWriter` configuration recorded in the MemWAL index.
+    pub writer_config_defaults: Option<HashMap<String, String>>,
+}
+
+impl TryFrom<LsmWriteSpec> for lancedb::table::LsmWriteSpec {
+    type Error = napi::Error;
+
+    fn try_from(value: LsmWriteSpec) -> napi::Result<Self> {
+        let maintained = value.maintained_indexes.unwrap_or_default();
+        let writer_config_defaults = value.writer_config_defaults.unwrap_or_default();
+        let spec = match value.spec_type.as_str() {
+            "bucket" => {
+                let column = value.column.ok_or_else(|| {
+                    napi::Error::from_reason("LsmWriteSpec bucket requires `column`")
+                })?;
+                let num_buckets = value.num_buckets.ok_or_else(|| {
+                    napi::Error::from_reason("LsmWriteSpec bucket requires `numBuckets`")
+                })?;
+                Self::bucket(column, num_buckets)
+            }
+            "identity" => {
+                let column = value.column.ok_or_else(|| {
+                    napi::Error::from_reason("LsmWriteSpec identity requires `column`")
+                })?;
+                Self::identity(column)
+            }
+            "unsharded" => Self::unsharded(),
+            other => {
+                return Err(napi::Error::from_reason(format!(
+                    "LsmWriteSpec `specType` must be 'bucket', 'identity', or 'unsharded', got '{}'",
+                    other
+                )));
+            }
+        };
+        Ok(spec
+            .with_maintained_indexes(maintained)
+            .with_writer_config_defaults(writer_config_defaults))
+    }
+}
+
 /// Statistics about a compaction operation.
 #[napi(object)]
 #[derive(Clone, Debug)]
@@ -572,6 +676,44 @@ pub struct OptimizeStats {
     pub prune: RemovalStats,
 }
 
+/// Progress snapshot for a write operation, delivered to the JS callback
+/// passed to `Table.add`.
+#[napi(object)]
+#[derive(Clone, Debug)]
+pub struct WriteProgressInfo {
+    /// Number of rows written so far.
+    pub output_rows: i64,
+    /// Number of bytes written so far.
+    pub output_bytes: i64,
+    /// Total rows expected, if the input source reports it.
+    /// Always set on the final callback (where `done` is `true`).
+    pub total_rows: Option<i64>,
+    /// Wall-clock seconds since monitoring started.
+    pub elapsed_seconds: f64,
+    /// Number of parallel write tasks currently in flight.
+    pub active_tasks: i64,
+    /// Total number of parallel write tasks (the write parallelism).
+    pub total_tasks: i64,
+    /// `true` for the final callback; `false` otherwise.
+    pub done: bool,
+}
+
+impl From<&lancedb::table::write_progress::WriteProgress> for WriteProgressInfo {
+    fn from(p: &lancedb::table::write_progress::WriteProgress) -> Self {
+        Self {
+            output_rows: p.output_rows() as i64,
+            output_bytes: p.output_bytes() as i64,
+            total_rows: p.total_rows().map(|n| n as i64),
+            elapsed_seconds: p.elapsed().as_secs_f64(),
+            active_tasks: p.active_tasks() as i64,
+            total_tasks: p.total_tasks() as i64,
+            done: p.done(),
+        }
+    }
+}
+
+type ProgressFn = ThreadsafeFunction<WriteProgressInfo, (), WriteProgressInfo, Status, false>;
+
 ///  A definition of a column alteration. The alteration changes the column at
 /// `path` to have the new name `name`, to be nullable if `nullable` is true,
 /// and to have the data type `data_type`. At least one of `rename` or `nullable`

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.31.0-beta.11"
+current_version = "0.32.1-beta.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/AGENTS.md

@@ -4,16 +4,26 @@ code is in the `src/` directory and the Python bindings are in the `lancedb/` di
 
 Common commands:
 
+* Bootstrap dev env: `uv run --extra tests --extra dev maturin develop --extras tests,dev`
 * Build: `make develop`
 * Format: `make format`
 * Lint: `make check`
 * Fix lints: `make fix`
-* Test: `make test`
-* Doc test: `make doctest`
+* Test: `uv run --extra tests pytest python/tests -vv --durations=10 -m "not slow and not s3_test"`
+* Run specific test: `uv run --extra tests pytest python/tests/<test_file>.py::<test_name> -q`
+* Doc test: `uv run --extra tests pytest --doctest-modules python/lancedb`
+
+Use the uv-managed environment declared by `uv.lock` for Python validation. Do
+not treat system `python`, global `pytest`, or missing editable-install errors
+as final blockers; bootstrap or enter the uv environment instead. `make test`
+and `make doctest` assume the development environment is already prepared.
 
 Before committing changes, run lints and then formatting.
 
-When you change the Rust code, you will need to recompile the Python bindings: `make develop`.
+When you change the Rust code, PyO3 binding code, or see a missing/stale
+`lancedb._lancedb`, recompile the Python bindings with
+`uv run --extra tests --extra dev maturin develop --extras tests,dev` before
+running tests.
 
 When you export new types from Rust to Python, you must manually update `python/lancedb/_lancedb.pyi`
 with the corresponding type hints. You can run `pyright` to check for type errors in the Python code.

python/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.31.0-beta.11"
+version = "0.32.1-beta.0"
 publish = false
 edition.workspace = true
 description = "Python bindings for LanceDB"
@@ -19,6 +19,7 @@ arrow = { version = "58.0.0", features = ["pyarrow"] }
 async-trait = "0.1"
 bytes = "1"
 lancedb = { path = "../rust/lancedb", default-features = false }
+datafusion-common.workspace = true
 lance-core.workspace = true
 lance-namespace.workspace = true
 lance-namespace-impls.workspace = true

python/python/lancedb/_lancedb.pyi

@@ -51,7 +51,7 @@ class PyExpr:
     def to_sql(self) -> str: ...
 
 def expr_col(name: str) -> PyExpr: ...
-def expr_lit(value: Union[bool, int, float, str]) -> PyExpr: ...
+def expr_lit(value: Union[bool, int, float, str, bytes]) -> PyExpr: ...
 def expr_func(name: str, args: List[PyExpr]) -> PyExpr: ...
 
 class Session:
@@ -217,6 +217,9 @@ class Table:
     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]]: ...
+    async def set_unenforced_primary_key(self, columns: List[str]) -> None: ...
+    async def set_lsm_write_spec(self, spec: LsmWriteSpec) -> None: ...
+    async def unset_lsm_write_spec(self) -> None: ...
     @property
     def tags(self) -> Tags: ...
     def query(self) -> Query: ...
@@ -255,6 +258,11 @@ class RecordBatchStream:
     def __aiter__(self) -> "RecordBatchStream": ...
     async def __anext__(self) -> pa.RecordBatch: ...
 
+class ColumnOrdering(TypedDict):
+    column_name: str
+    ascending: bool
+    nulls_first: bool
+
 class Query:
     def where(self, filter: str): ...
     def where_expr(self, expr: PyExpr): ...
@@ -268,6 +276,7 @@ class Query:
     def postfilter(self): ...
     def nearest_to(self, query_vec: pa.Array) -> VectorQuery: ...
     def nearest_to_text(self, query: dict) -> FTSQuery: ...
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]): ...
     async def output_schema(self) -> pa.Schema: ...
     async def execute(
         self, max_batch_length: Optional[int], timeout: Optional[timedelta]
@@ -296,6 +305,7 @@ class FTSQuery:
     def get_query(self) -> str: ...
     def add_query_vector(self, query_vec: pa.Array) -> None: ...
     def nearest_to(self, query_vec: pa.Array) -> HybridQuery: ...
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]): ...
     async def output_schema(self) -> pa.Schema: ...
     async def execute(
         self, max_batch_length: Optional[int], timeout: Optional[timedelta]
@@ -321,6 +331,7 @@ class VectorQuery:
     def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
     def nearest_to_text(self, query: dict) -> HybridQuery: ...
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]): ...
     def to_query_request(self) -> PyQueryRequest: ...
 
 class HybridQuery:
@@ -339,6 +350,7 @@ class HybridQuery:
     def minimum_nprobes(self, minimum_nprobes: int): ...
     def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]): ...
     def to_vector_query(self) -> VectorQuery: ...
     def to_fts_query(self) -> FTSQuery: ...
     def get_limit(self) -> int: ...
@@ -368,6 +380,7 @@ class PyQueryRequest:
     bypass_vector_index: Optional[bool]
     postfilter: Optional[bool]
     norm: Optional[str]
+    order_by: Optional[List[ColumnOrdering]]
 
 class CompactionStats:
     fragments_removed: int
@@ -408,6 +421,37 @@ class MergeResult:
     num_deleted_rows: int
     num_attempts: int
 
+class LsmWriteSpec:
+    """Specification selecting Lance's MemWAL LSM-style write path for
+    `merge_insert`."""
+
+    @staticmethod
+    def bucket(column: str, num_buckets: int) -> "LsmWriteSpec": ...
+    @staticmethod
+    def identity(column: str) -> "LsmWriteSpec": ...
+    @staticmethod
+    def unsharded() -> "LsmWriteSpec": ...
+    def with_maintained_indexes(self, indexes: List[str]) -> "LsmWriteSpec":
+        """Return a copy of this spec asking the MemWAL to keep the named
+        indexes up to date as rows are appended."""
+        ...
+    def with_writer_config_defaults(self, defaults: Dict[str, str]) -> "LsmWriteSpec":
+        """Return a copy of this spec recording the given default
+        `ShardWriter` configuration in the MemWAL index."""
+        ...
+    @property
+    def spec_type(self) -> str:
+        """One of 'bucket', 'identity', or 'unsharded'."""
+        ...
+    @property
+    def column(self) -> Optional[str]: ...
+    @property
+    def num_buckets(self) -> Optional[int]: ...
+    @property
+    def maintained_indexes(self) -> List[str]: ...
+    @property
+    def writer_config_defaults(self) -> Dict[str, str]: ...
+
 class AddColumnsResult:
     version: int
 

python/python/lancedb/expr.py

@@ -63,7 +63,7 @@ def _coerce(value: "ExprLike") -> "Expr":
 
 
 # Type alias used in annotations.
-ExprLike = Union["Expr", bool, int, float, str]
+ExprLike = Union["Expr", bool, int, float, str, bytes]
 
 
 class Expr:
@@ -261,13 +261,13 @@ def col(name: str) -> Expr:
     return Expr(expr_col(name))
 
 
-def lit(value: Union[bool, int, float, str]) -> Expr:
+def lit(value: Union[bool, int, float, str, bytes]) -> Expr:
     """Create a literal (constant) value expression.
 
     Parameters
     ----------
     value:
-        A Python ``bool``, ``int``, ``float``, or ``str``.
+        A Python ``bool``, ``int``, ``float``, ``str``, or ``bytes``.
 
     Examples
     --------

python/python/lancedb/namespace_utils.py

@@ -6,22 +6,44 @@
 from typing import Optional
 
 
+_CREATE_NAMESPACE_MODES = frozenset({"create", "exist_ok", "overwrite"})
+_DROP_NAMESPACE_MODES = frozenset({"SKIP", "FAIL"})
+_DROP_NAMESPACE_BEHAVIORS = frozenset({"RESTRICT", "CASCADE"})
+
+
 def _normalize_create_namespace_mode(mode: Optional[str]) -> Optional[str]:
     """Normalize create namespace mode to lowercase (API expects lowercase)."""
     if mode is None:
         return None
-    return mode.lower()
+    normalized = mode.lower()
+    if normalized not in _CREATE_NAMESPACE_MODES:
+        raise ValueError(
+            f"Invalid create namespace mode {mode!r}: "
+            f"expected one of 'create', 'exist_ok', 'overwrite'"
+        )
+    return normalized
 
 
 def _normalize_drop_namespace_mode(mode: Optional[str]) -> Optional[str]:
     """Normalize drop namespace mode to uppercase (API expects uppercase)."""
     if mode is None:
         return None
-    return mode.upper()
+    normalized = mode.upper()
+    if normalized not in _DROP_NAMESPACE_MODES:
+        raise ValueError(
+            f"Invalid drop namespace mode {mode!r}: expected one of 'skip', 'fail'"
+        )
+    return normalized
 
 
 def _normalize_drop_namespace_behavior(behavior: Optional[str]) -> Optional[str]:
     """Normalize drop namespace behavior to uppercase (API expects uppercase)."""
     if behavior is None:
         return None
-    return behavior.upper()
+    normalized = behavior.upper()
+    if normalized not in _DROP_NAMESPACE_BEHAVIORS:
+        raise ValueError(
+            f"Invalid drop namespace behavior {behavior!r}: "
+            f"expected one of 'restrict', 'cascade'"
+        )
+    return normalized

python/python/lancedb/permutation.py

@@ -968,22 +968,32 @@ class Permutation:
         new.transform_fn = transform
         return new
 
+    def take_offsets(self, offsets: list[int]) -> Any:
+        """
+        Take rows from the permutation by offset
+
+        The returned value is passed through the permutation's current transform,
+        so `with_format` and `with_transform` affect this method in the same way
+        they affect iteration.
+        """
+
+        async def do_take_offsets():
+            return await self.reader.take_offsets(offsets, selection=self.selection)
+
+        batch = LOOP.run(do_take_offsets())
+        return self.transform_fn(batch)
+
     def __getitem__(self, index: int) -> Any:
         """
         Returns a single row from the permutation by offset
         """
-        return self.__getitems__([index])
+        return self.take_offsets([index])
 
     def __getitems__(self, indices: list[int]) -> Any:
         """
         Returns rows from the permutation by offset
         """
-
-        async def do_getitems():
-            return await self.reader.take_offsets(indices, selection=self.selection)
-
-        batch = LOOP.run(do_getitems())
-        return self.transform_fn(batch)
+        return self.take_offsets(indices)
 
     @deprecated(details="Use with_skip instead")
     def skip(self, skip: int) -> "Permutation":

python/python/lancedb/query.py

@@ -92,6 +92,12 @@ def ensure_vector_query(
         return val
 
 
+class ColumnOrdering(pydantic.BaseModel):
+    column_name: str
+    ascending: bool = True
+    nulls_first: bool = False
+
+
 class FullTextQueryType(str, Enum):
     MATCH = "match"
     MATCH_PHRASE = "match_phrase"
@@ -504,6 +510,8 @@ class Query(pydantic.BaseModel):
     # Bypass the vector index and use a brute force search
     bypass_vector_index: Optional[bool] = None
 
+    order_by: Optional[List[ColumnOrdering]] = None
+
     @classmethod
     def from_inner(cls, req: PyQueryRequest) -> Self:
         query = cls()
@@ -524,6 +532,8 @@ class Query(pydantic.BaseModel):
         query.refine_factor = req.refine_factor
         query.bypass_vector_index = req.bypass_vector_index
         query.postfilter = req.postfilter
+        if req.order_by is not None:
+            query.order_by = [ColumnOrdering(**o) for o in req.order_by]
         if req.full_text_search is not None:
             query.full_text_query = FullTextSearchQuery(
                 columns=None,
@@ -572,9 +582,22 @@ class LanceQueryBuilder(ABC):
             If "auto", the query type is inferred based on the query.
         vector_column_name: str
             The name of the vector column to use for vector search.
+        ordering_field_name: Optional[str]
+            .. deprecated:: 0.27.0
+                Use ``order_by()`` method instead.
+        fts_columns: Optional[Union[str, List[str]]]
+            The columns to search in for full text search.
         fast_search: bool
             Skip flat search of unindexed data.
         """
+        if ordering_field_name is not None:
+            import warnings
+
+            warnings.warn(
+                "ordering_field_name is deprecated, use .order_by() method instead.",
+                DeprecationWarning,
+                stacklevel=2,
+            )
         # Check hybrid search first as it supports empty query pattern
         if query_type == "hybrid":
             # hybrid fts and vector query
@@ -671,6 +694,7 @@ class LanceQueryBuilder(ABC):
         self._text = None
         self._ef = None
         self._bypass_vector_index = None
+        self._order_by = None
 
     @deprecation.deprecated(
         deprecated_in="0.3.1",
@@ -694,6 +718,7 @@ class LanceQueryBuilder(ABC):
         flatten: Optional[Union[int, bool]] = None,
         *,
         timeout: Optional[timedelta] = None,
+        **kwargs,
     ) -> "pd.DataFrame":
         """
         Execute the query and return the results as a pandas DataFrame.
@@ -711,9 +736,12 @@ class LanceQueryBuilder(ABC):
         timeout: Optional[timedelta]
             The maximum time to wait for the query to complete.
             If None, wait indefinitely.
+        **kwargs
+            Forwarded to pyarrow.Table.to_pandas after query execution and
+            optional flattening.
         """
         tbl = flatten_columns(self.to_arrow(timeout=timeout), flatten)
-        return tbl.to_pandas()
+        return tbl.to_pandas(**kwargs)
 
     @abstractmethod
     def to_arrow(self, *, timeout: Optional[timedelta] = None) -> pa.Table:
@@ -947,6 +975,24 @@ class LanceQueryBuilder(ABC):
         """  # noqa: E501
         return self._table._explain_plan(self.to_query_object(), verbose=verbose)
 
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]) -> Self:
+        """
+        Set the ordering for the results.
+
+        Parameters
+        ----------
+        ordering: Optional[List[ColumnOrdering]]
+            The ordering to use for the results.  If None, then the default ordering
+            will be used.
+
+        Returns
+        -------
+        LanceQueryBuilder
+            The LanceQueryBuilder object.
+        """
+        self._order_by = ordering
+        return self
+
     def analyze_plan(self) -> str:
         """
         Run the query and return its execution plan with runtime metrics.
@@ -1314,6 +1360,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
             fast_search=self._fast_search,
             ef=self._ef,
             bypass_vector_index=self._bypass_vector_index,
+            order_by=self._order_by,
         )
 
     def to_batches(
@@ -1465,7 +1512,9 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
         super().__init__(table)
         self._query = query
         self._phrase_query = False
-        self.ordering_field_name = ordering_field_name
+        # Deprecated compatibility parameter. Native FTS ordering is now
+        # configured through order_by(); LanceQueryBuilder.create emits the warning.
+        _ = ordering_field_name
         self._reranker = None
         self._fast_search = fast_search
         if isinstance(fts_columns, str):
@@ -1514,6 +1563,7 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
             ),
             offset=self._offset,
             fast_search=self._fast_search,
+            order_by=self._order_by,
         )
 
     def output_schema(self) -> pa.Schema:
@@ -1579,6 +1629,7 @@ class LanceEmptyQueryBuilder(LanceQueryBuilder):
             limit=self._limit,
             with_row_id=self._with_row_id,
             offset=self._offset,
+            order_by=self._order_by,
         )
 
     def output_schema(self) -> pa.Schema:
@@ -2305,6 +2356,7 @@ class AsyncQueryBase(object):
         self,
         flatten: Optional[Union[int, bool]] = None,
         timeout: Optional[timedelta] = None,
+        **kwargs,
     ) -> "pd.DataFrame":
         """
         Execute the query and collect the results into a pandas DataFrame.
@@ -2337,10 +2389,13 @@ class AsyncQueryBase(object):
             The maximum time to wait for the query to complete.
             If not specified, no timeout is applied. If the query does not
             complete within the specified time, an error will be raised.
+        **kwargs
+            Forwarded to pyarrow.Table.to_pandas after query execution and
+            optional flattening.
         """
         return (
             flatten_columns(await self.to_arrow(timeout=timeout), flatten)
-        ).to_pandas()
+        ).to_pandas(**kwargs)
 
     async def to_polars(
         self,
@@ -2502,6 +2557,27 @@ class AsyncStandardQuery(AsyncQueryBase):
         self._inner.offset(offset)
         return self
 
+    def order_by(self, ordering: Optional[List[ColumnOrdering]]) -> Self:
+        """
+        Set the ordering for the results.
+
+        Parameters
+        ----------
+        ordering: Optional[List[ColumnOrdering]]
+            The ordering to use for the results.  If None, then the default ordering
+            will be used.
+        """
+        if ordering is None:
+            self._inner.order_by(None)
+        else:
+            self._inner.order_by(
+                [
+                    o.model_dump() if hasattr(o, "model_dump") else o.dict()
+                    for o in ordering
+                ]
+            )
+        return self
+
     def fast_search(self) -> Self:
         """
         Skip searching un-indexed data.
@@ -3321,6 +3397,7 @@ class BaseQueryBuilder(object):
         self,
         flatten: Optional[Union[int, bool]] = None,
         timeout: Optional[timedelta] = None,
+        **kwargs,
     ) -> "pd.DataFrame":
         """
         Execute the query and collect the results into a pandas DataFrame.
@@ -3353,8 +3430,11 @@ class BaseQueryBuilder(object):
             The maximum time to wait for the query to complete.
             If not specified, no timeout is applied. If the query does not
             complete within the specified time, an error will be raised.
+        **kwargs
+            Forwarded to pyarrow.Table.to_pandas after query execution and
+            optional flattening.
         """
-        return LOOP.run(self._inner.to_pandas(flatten, timeout))
+        return LOOP.run(self._inner.to_pandas(flatten, timeout, **kwargs))
 
     def to_polars(
         self,

python/python/lancedb/remote/table.py

@@ -14,6 +14,7 @@ from lancedb._lancedb import (
     DeleteResult,
     DropColumnsResult,
     IndexConfig,
+    LsmWriteSpec,
     MergeResult,
     UpdateResult,
 )
@@ -39,7 +40,7 @@ from lancedb.embeddings import EmbeddingFunctionRegistry
 from lancedb.table import _normalize_progress
 
 from ..query import LanceVectorQueryBuilder, LanceQueryBuilder, LanceTakeQueryBuilder
-from ..table import AsyncTable, IndexStatistics, Query, Table, Tags
+from ..table import AsyncTable, BlobMode, IndexStatistics, Query, Table, Tags
 from ..types import BaseTokenizerType
 
 
@@ -100,7 +101,7 @@ class RemoteTable(Table):
         """to_arrow() is not yet supported on LanceDB cloud."""
         raise NotImplementedError("to_arrow() is not yet supported on LanceDB cloud.")
 
-    def to_pandas(self):
+    def to_pandas(self, blob_mode: BlobMode = "lazy", **kwargs):
         """to_pandas() is not yet supported on LanceDB cloud."""
         raise NotImplementedError("to_pandas() is not yet supported on LanceDB cloud.")
 
@@ -655,6 +656,18 @@ class RemoteTable(Table):
     def drop_columns(self, columns: Iterable[str]) -> DropColumnsResult:
         return LOOP.run(self._table.drop_columns(columns))
 
+    def set_unenforced_primary_key(self, columns: Union[str, Iterable[str]]) -> None:
+        """Not supported on LanceDB Cloud."""
+        return LOOP.run(self._table.set_unenforced_primary_key(columns))
+
+    def set_lsm_write_spec(self, spec: "LsmWriteSpec") -> None:
+        """Not supported on LanceDB Cloud."""
+        return LOOP.run(self._table.set_lsm_write_spec(spec))
+
+    def unset_lsm_write_spec(self) -> None:
+        """Not supported on LanceDB Cloud."""
+        return LOOP.run(self._table.unset_lsm_write_spec())
+
     def drop_index(self, index_name: str):
         return LOOP.run(self._table.drop_index(index_name))
 

python/python/lancedb/table.py

@@ -87,6 +87,8 @@ from .util import (
 )
 from .index import lang_mapping
 
+BlobMode = Literal["lazy", "bytes", "descriptions"]
+
 _MODEL_BACKED_TOKENIZER_PREFIXES = ("jieba", "lindera")
 _MODEL_BACKED_TOKENIZER_ERRORS = (
     "unknown base tokenizer",
@@ -154,6 +156,7 @@ if TYPE_CHECKING:
         AlterColumnsResult,
         DeleteResult,
         DropColumnsResult,
+        LsmWriteSpec,
         MergeResult,
         UpdateResult,
     )
@@ -759,14 +762,22 @@ class Table(ABC):
         """
         raise NotImplementedError
 
-    def to_pandas(self) -> "pandas.DataFrame":
+    def to_pandas(self, blob_mode: BlobMode = "lazy", **kwargs) -> "pandas.DataFrame":
         """Return the table as a pandas DataFrame.
 
+        Parameters
+        ----------
+        blob_mode: str, default "lazy"
+            Controls how blob columns are returned for backends that support
+            Lance blob-aware pandas conversion.
+        **kwargs
+            Forwarded to PyArrow / Lance pandas conversion.
+
         Returns
         -------
         pd.DataFrame
         """
-        return self.to_arrow().to_pandas()
+        return self.to_arrow().to_pandas(**kwargs)
 
     @abstractmethod
     def to_arrow(self) -> pa.Table:
@@ -2182,14 +2193,27 @@ class LanceTable(Table):
         """Return the first n rows of the table."""
         return LOOP.run(self._table.head(n))
 
-    def to_pandas(self) -> "pd.DataFrame":
+    def to_pandas(self, blob_mode: BlobMode = "lazy", **kwargs) -> "pd.DataFrame":
         """Return the table as a pandas DataFrame.
 
+        Parameters
+        ----------
+        blob_mode: str, default "lazy"
+            Controls how Lance blob columns are returned.
+        **kwargs
+            Forwarded to Lance pandas conversion.
+
         Returns
         -------
         pd.DataFrame
         """
-        return self.to_arrow().to_pandas()
+        if blob_mode == "lazy" and (
+            self._namespace_client is not None
+            or get_uri_scheme(self._dataset_path) == "memory"
+        ):
+            return self.to_arrow().to_pandas(**kwargs)
+
+        return self.to_lance().to_pandas(blob_mode=blob_mode, **kwargs)
 
     def to_arrow(self) -> pa.Table:
         """Return the table as a pyarrow Table.
@@ -2518,11 +2542,6 @@ class LanceTable(Table):
                 "at a time. To search over multiple text fields, create a "
                 "separate FTS index for each field."
             )
-        if "." in field_names:
-            raise ValueError(
-                "Native FTS indexes can only be created on top-level fields. "
-                f"Received nested field path: {field_names!r}."
-            )
 
         if tokenizer_name is None:
             tokenizer_configs = {
@@ -3263,6 +3282,21 @@ class LanceTable(Table):
     def drop_columns(self, columns: Iterable[str]) -> DropColumnsResult:
         return LOOP.run(self._table.drop_columns(columns))
 
+    def set_unenforced_primary_key(self, columns: Union[str, Iterable[str]]) -> None:
+        """Set the unenforced primary key. See
+        [`AsyncTable.set_unenforced_primary_key`][lancedb.AsyncTable.set_unenforced_primary_key]."""
+        return LOOP.run(self._table.set_unenforced_primary_key(columns))
+
+    def set_lsm_write_spec(self, spec: "LsmWriteSpec") -> None:
+        """Install an LsmWriteSpec. See
+        [`AsyncTable.set_lsm_write_spec`][lancedb.AsyncTable.set_lsm_write_spec]."""
+        return LOOP.run(self._table.set_lsm_write_spec(spec))
+
+    def unset_lsm_write_spec(self) -> None:
+        """Remove the LsmWriteSpec. See
+        [`AsyncTable.unset_lsm_write_spec`][lancedb.AsyncTable.unset_lsm_write_spec]."""
+        return LOOP.run(self._table.unset_lsm_write_spec())
+
     def uses_v2_manifest_paths(self) -> bool:
         """
         Check if the table is using the new v2 manifest paths.
@@ -3808,6 +3842,69 @@ class AsyncTable:
         Any attempt to use the table after it has been closed will raise an error."""
         return self._inner.close()
 
+    async def set_unenforced_primary_key(
+        self, columns: Union[str, Iterable[str]]
+    ) -> None:
+        """Set the unenforced primary key for this table to the given
+        ordered list of columns.
+
+        "Unenforced" means LanceDB does not check uniqueness on writes; the
+        columns are recorded in the schema as the primary key so that
+        features such as `merge_insert` can use them. Calling this again
+        replaces any previously-set primary key.
+
+        Parameters
+        ----------
+        columns : str or Iterable[str]
+            Either a single column name (single-column key) or an ordered
+            iterable of column names (composite key). Each column dtype
+            must be one of: int32, int64, utf8, large_utf8, binary,
+            large_binary, fixed_size_binary.
+        """
+        if isinstance(columns, str):
+            columns = [columns]
+        else:
+            columns = list(columns)
+        await self._inner.set_unenforced_primary_key(columns)
+
+    async def set_lsm_write_spec(self, spec: "LsmWriteSpec") -> None:
+        """Install an LsmWriteSpec on this table.
+
+        The spec selects Lance's MemWAL LSM-style write path for future
+        `merge_insert` calls. ``LsmWriteSpec`` chooses one of three sharding
+        strategies:
+
+        - ``LsmWriteSpec.bucket(column, num_buckets)`` — hash-bucket writes by
+          the single-column unenforced primary key.
+        - ``LsmWriteSpec.identity(column)`` — shard by the raw value of a
+          scalar column.
+        - ``LsmWriteSpec.unsharded()`` — route every write to a single shard.
+
+        All variants require the table to have an unenforced primary key set
+        via [`set_unenforced_primary_key`]; bucket sharding additionally
+        requires it to be the single column being bucketed.
+
+        Parameters
+        ----------
+        spec : LsmWriteSpec
+            The sharding spec to install.
+
+        Examples
+        --------
+        >>> from lancedb._lancedb import LsmWriteSpec
+        >>> # table.set_unenforced_primary_key("id")
+        >>> # table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 16))
+        """
+        await self._inner.set_lsm_write_spec(spec)
+
+    async def unset_lsm_write_spec(self) -> None:
+        """Remove the LsmWriteSpec from this table.
+
+        Reverts to the standard `merge_insert` write path. Errors if no spec
+        is currently set.
+        """
+        await self._inner.unset_lsm_write_spec()
+
     @property
     def name(self) -> str:
         """The name of the table."""
@@ -3866,14 +3963,39 @@ class AsyncTable:
         """
         return AsyncQuery(self._inner.query())
 
-    async def to_pandas(self) -> "pd.DataFrame":
+    async def _to_lance(self, **kwargs) -> lance.LanceDataset:
+        try:
+            import lance
+        except ImportError:
+            raise ImportError(
+                "The lance library is required to use this function. "
+                "Please install with `pip install pylance`."
+            )
+
+        return lance.dataset(
+            await self.uri(),
+            version=await self.version(),
+            storage_options=await self.latest_storage_options(),
+            **kwargs,
+        )
+
+    async def to_pandas(self, blob_mode: BlobMode = "lazy", **kwargs) -> "pd.DataFrame":
         """Return the table as a pandas DataFrame.
 
+        Parameters
+        ----------
+        blob_mode: str, default "lazy"
+            Controls how Lance blob columns are returned.
+        **kwargs
+            Forwarded to PyArrow / Lance pandas conversion.
+
         Returns
         -------
         pd.DataFrame
         """
-        return (await self.to_arrow()).to_pandas()
+        if blob_mode == "lazy":
+            return (await self.to_arrow()).to_pandas(**kwargs)
+        return (await self._to_lance()).to_pandas(blob_mode=blob_mode, **kwargs)
 
     async def to_arrow(self) -> pa.Table:
         """Return the table as a pyarrow Table.
@@ -4512,6 +4634,8 @@ class AsyncTable:
             async_query = async_query.fast_search()
         if query.with_row_id:
             async_query = async_query.with_row_id()
+        if query.order_by:
+            async_query = async_query.order_by(query.order_by)
 
         if query.vector:
             async_query = async_query.nearest_to(query.vector).distance_range(

python/python/tests/models/lindera/ipadic/config.yml

@@ -1,4 +1,3 @@
 segmenter:
   mode: "normal"
-  dictionary:
-    path: "./python/tests/models/lindera/ipadic/main"
+  dictionary: "./python/tests/models/lindera/ipadic/main"

python/python/tests/test_db.py

@@ -914,6 +914,29 @@ def test_local_namespace_operations(tmp_path):
     assert db.list_namespaces().namespaces == []
 
 
+def test_create_namespace_invalid_mode_raises(tmp_path):
+    """Unrecognized create namespace modes raise a clear error."""
+    db = lancedb.connect(tmp_path)
+    with pytest.raises(ValueError, match="Invalid create namespace mode"):
+        db.create_namespace(["child"], mode="frobnicate")
+
+
+def test_drop_namespace_invalid_mode_raises(tmp_path):
+    """Unrecognized drop namespace modes raise a clear error."""
+    db = lancedb.connect(tmp_path)
+    db.create_namespace(["child"])
+    with pytest.raises(ValueError, match="Invalid drop namespace mode"):
+        db.drop_namespace(["child"], mode="frobnicate")
+
+
+def test_drop_namespace_invalid_behavior_raises(tmp_path):
+    """Unrecognized drop namespace behaviors raise a clear error."""
+    db = lancedb.connect(tmp_path)
+    db.create_namespace(["child"])
+    with pytest.raises(ValueError, match="Invalid drop namespace behavior"):
+        db.drop_namespace(["child"], behavior="frobnicate")
+
+
 def test_clone_table_latest_version(tmp_path):
     """Test cloning a table with the latest version (default behavior)"""
     import os

python/python/tests/test_fts.py

@@ -29,6 +29,7 @@ from lancedb.query import (
     MultiMatchQuery,
     PhraseQuery,
     BooleanQuery,
+    ColumnOrdering,
     Occur,
     LanceFtsQueryBuilder,
 )
@@ -116,8 +117,7 @@ def lindera_ipadic(language_model_home):
     config_path.write_text(
         "segmenter:\n"
         '  mode: "normal"\n'
-        "  dictionary:\n"
-        f'    path: "{extracted_model.resolve().as_posix()}"\n',
+        f'  dictionary: "{extracted_model.resolve().as_posix()}"\n',
         encoding="utf-8",
     )
 
@@ -500,6 +500,36 @@ async def test_search_fts_specify_column_async(async_table):
         pass
 
 
+def test_search_order_by_descending(table):
+    table.create_fts_index("text")
+    rows = (
+        table.search("puppy")
+        .order_by([ColumnOrdering(column_name="count", ascending=False)])
+        .limit(20)
+        .select(["text", "count"])
+        .to_list()
+    )
+
+    for r in rows:
+        assert "puppy" in r["text"]
+    assert sorted(rows, key=lambda x: x["count"], reverse=True) == rows
+
+
+def test_search_order_by_ascending(table):
+    table.create_fts_index("text")
+    rows = (
+        table.search("puppy")
+        .order_by([ColumnOrdering(column_name="count", ascending=True)])
+        .limit(20)
+        .select(["text", "count"])
+        .to_list()
+    )
+
+    for r in rows:
+        assert "puppy" in r["text"]
+    assert sorted(rows, key=lambda x: x["count"]) == rows
+
+
 def test_create_index_from_table(tmp_path, table):
     table.create_fts_index("text")
     df = table.search("puppy").limit(5).select(["text"]).to_pandas()
@@ -533,8 +563,19 @@ def test_create_index_multiple_columns(tmp_path, table):
 
 
 def test_nested_schema(tmp_path, table):
-    with pytest.raises(ValueError, match="top-level fields"):
-        table.create_fts_index("nested.text")
+    table.create_fts_index("nested.text")
+    indices = table.list_indices()
+    assert len(indices) == 1
+    assert indices[0].index_type == "FTS"
+    assert indices[0].columns == ["nested.text"]
+
+    results = (
+        table.search("puppy", query_type="fts", fts_columns="nested.text")
+        .limit(5)
+        .to_list()
+    )
+    assert len(results) > 0
+    assert all("puppy" in row["nested"]["text"] for row in results)
 
 
 def test_search_index_with_filter(table):

python/python/tests/test_lsm_write_spec.py

@@ -0,0 +1,149 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+"""Tests for installing and clearing an LsmWriteSpec via
+`Table.set_lsm_write_spec` / `Table.unset_lsm_write_spec`.
+"""
+
+from datetime import timedelta
+
+import lancedb
+import pyarrow as pa
+import pytest
+from lancedb._lancedb import LsmWriteSpec
+
+SCHEMA = pa.schema(
+    [
+        pa.field("id", pa.utf8(), nullable=False),
+        pa.field("v", pa.int32(), nullable=False),
+    ]
+)
+
+
+def _batch(ids, vs):
+    return pa.RecordBatch.from_arrays(
+        [pa.array(ids, type=pa.utf8()), pa.array(vs, type=pa.int32())],
+        schema=SCHEMA,
+    )
+
+
+def _reader(ids, vs):
+    return pa.RecordBatchReader.from_batches(SCHEMA, [_batch(ids, vs)])
+
+
+def _make_table(tmp_path):
+    db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
+    table = db.create_table("t", _reader(["seed"], [0]))
+    return db, table
+
+
+def test_set_lsm_write_spec_validates(tmp_path):
+    _db, table = _make_table(tmp_path)
+
+    # No PK set yet.
+    with pytest.raises(Exception, match="primary key"):
+        table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 4))
+
+    table.set_unenforced_primary_key("id")
+
+    # Column mismatch.
+    with pytest.raises(Exception, match="match"):
+        table.set_lsm_write_spec(LsmWriteSpec.bucket("v", 4))
+
+    # Out-of-range num_buckets.
+    with pytest.raises(Exception, match="num_buckets"):
+        table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 0))
+    with pytest.raises(Exception, match="num_buckets"):
+        table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 1025))
+
+    # Happy path then mutation rejected.
+    table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 4))
+    with pytest.raises(Exception, match="mutation"):
+        table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 8))
+
+
+def test_unset_lsm_write_spec(tmp_path):
+    _db, table = _make_table(tmp_path)
+
+    # unset errors when no spec is set.
+    with pytest.raises(Exception, match="no LSM write spec"):
+        table.unset_lsm_write_spec()
+
+    # Install a spec, then remove it; afterwards a fresh spec can be set.
+    table.set_unenforced_primary_key("id")
+    table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 4))
+    table.unset_lsm_write_spec()
+    # A second unset errors — there is no spec left to remove.
+    with pytest.raises(Exception, match="no LSM write spec"):
+        table.unset_lsm_write_spec()
+    table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 8))
+
+
+def test_set_unsharded_spec(tmp_path):
+    _db, table = _make_table(tmp_path)
+    # Lance MemWAL still requires a primary key on the dataset; Unsharded
+    # just skips per-row hashing.
+    table.set_unenforced_primary_key("id")
+    table.set_lsm_write_spec(LsmWriteSpec.unsharded())
+    table.unset_lsm_write_spec()
+
+
+def test_lsm_write_spec_repr():
+    s = LsmWriteSpec.bucket("id", 4)
+    assert s.spec_type == "bucket"
+    assert s.column == "id"
+    assert s.num_buckets == 4
+    assert s.maintained_indexes == []
+    assert "bucket" in repr(s)
+    assert "id" in repr(s)
+    assert "4" in repr(s)
+
+    u = LsmWriteSpec.unsharded()
+    assert u.spec_type == "unsharded"
+    assert u.column is None
+    assert u.num_buckets is None
+    assert "unsharded" in repr(u)
+
+
+def test_lsm_write_spec_with_maintained_indexes():
+    s = LsmWriteSpec.bucket("id", 4).with_maintained_indexes(["idx_a", "idx_b"])
+    assert s.maintained_indexes == ["idx_a", "idx_b"]
+
+
+@pytest.mark.asyncio
+async def test_async_set_unset_lsm_write_spec(tmp_path):
+    db = await lancedb.connect_async(
+        tmp_path, read_consistency_interval=timedelta(seconds=0)
+    )
+    table = await db.create_table(
+        "t",
+        pa.RecordBatchReader.from_batches(SCHEMA, [_batch(["seed"], [0])]),
+    )
+
+    await table.set_unenforced_primary_key("id")
+    await table.set_lsm_write_spec(LsmWriteSpec.bucket("id", 4))
+    await table.unset_lsm_write_spec()
+    # A second unset errors.
+    with pytest.raises(Exception, match="no LSM write spec"):
+        await table.unset_lsm_write_spec()
+
+
+def test_set_identity_spec(tmp_path):
+    _db, table = _make_table(tmp_path)
+    # Identity sharding still requires an unenforced primary key on the
+    # table; it shards by the raw value of the given column.
+    table.set_unenforced_primary_key("id")
+    table.set_lsm_write_spec(LsmWriteSpec.identity("v"))
+    table.unset_lsm_write_spec()
+
+
+def test_lsm_write_spec_identity_and_writer_config_defaults():
+    s = LsmWriteSpec.identity("v")
+    assert s.spec_type == "identity"
+    assert s.column == "v"
+    assert s.num_buckets is None
+    assert "identity" in repr(s)
+
+    s = s.with_writer_config_defaults({"durable_write": "false"})
+    assert s.writer_config_defaults == {"durable_write": "false"}
+    assert "durable_write" in repr(s)

python/python/tests/test_permutation.py

@@ -1080,3 +1080,29 @@ def test_getitems_invalid_offset(some_permutation: Permutation):
     """Test __getitems__ with an out-of-range offset raises an error."""
     with pytest.raises(Exception):
         some_permutation.__getitems__([999999])
+
+
+def test_take_offsets(some_permutation: Permutation):
+    result = some_permutation.take_offsets([0, 1, 2])
+
+    assert isinstance(result, list)
+    assert "id" in result[0]
+    assert "value" in result[0]
+    assert len(result) == 3
+
+
+def test_take_offsets_empty_identity_permutation(mem_db):
+    tbl = mem_db.create_table(
+        "test_table", pa.table({"id": range(10), "value": range(10)})
+    )
+    permutation = Permutation.identity(tbl)
+
+    result = permutation.take_offsets([])
+
+    assert result == []
+
+
+def test_take_offsets_empty_permutation(some_permutation: Permutation):
+    result = some_permutation.take_offsets([])
+
+    assert result == []

python/python/tests/test_primary_key.py

@@ -0,0 +1,79 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+"""Tests for Table.set_unenforced_primary_key."""
+
+from datetime import timedelta
+
+import lancedb
+import pyarrow as pa
+import pytest
+
+
+def _empty_table(path, schema):
+    db = lancedb.connect(path, read_consistency_interval=timedelta(seconds=0))
+    return db.create_table("t", schema=schema)
+
+
+def test_set_unenforced_primary_key_accepts_string_or_one_element_list(tmp_path):
+    schema = pa.schema([pa.field("id", pa.int64(), nullable=False)])
+
+    # Bare string.
+    table = _empty_table(tmp_path / "s", schema)
+    table.set_unenforced_primary_key("id")
+
+    # One-element list.
+    table = _empty_table(tmp_path / "l", schema)
+    table.set_unenforced_primary_key(["id"])
+
+
+def test_set_unenforced_primary_key_rejects_compound_and_empty(tmp_path):
+    table = _empty_table(
+        tmp_path,
+        pa.schema(
+            [
+                pa.field("a", pa.utf8(), nullable=False),
+                pa.field("b", pa.int64(), nullable=False),
+            ]
+        ),
+    )
+    # Compound keys are not supported.
+    with pytest.raises(Exception, match="compound"):
+        table.set_unenforced_primary_key(["a", "b"])
+    # Empty input.
+    with pytest.raises(Exception, match="required"):
+        table.set_unenforced_primary_key([])
+
+
+def test_set_unenforced_primary_key_is_immutable(tmp_path):
+    table = _empty_table(
+        tmp_path,
+        pa.schema(
+            [
+                pa.field("a", pa.utf8(), nullable=False),
+                pa.field("b", pa.int64(), nullable=False),
+            ]
+        ),
+    )
+    table.set_unenforced_primary_key("a")
+    # The primary key cannot be changed or re-set once installed.
+    with pytest.raises(Exception, match="already set"):
+        table.set_unenforced_primary_key("b")
+    with pytest.raises(Exception, match="already set"):
+        table.set_unenforced_primary_key("a")
+
+
+def test_set_unenforced_primary_key_validates(tmp_path):
+    table = _empty_table(
+        tmp_path / "t", pa.schema([pa.field("id", pa.utf8(), nullable=False)])
+    )
+    # Unknown column.
+    with pytest.raises(Exception, match="not found"):
+        table.set_unenforced_primary_key("nonexistent")
+
+    # Unsupported dtype (Float32 not in the supported set).
+    bad = _empty_table(
+        tmp_path / "bad", pa.schema([pa.field("id", pa.float32(), nullable=False)])
+    )
+    with pytest.raises(Exception, match="not supported"):
+        bad.set_unenforced_primary_key("id")

python/python/tests/test_query.py

@@ -25,6 +25,7 @@ from lancedb.query import (
     AsyncHybridQuery,
     AsyncQueryBase,
     AsyncVectorQuery,
+    ColumnOrdering,
     LanceVectorQueryBuilder,
     MatchQuery,
     PhraseQuery,
@@ -164,6 +165,87 @@ def test_offset(table):
     assert len(results_with_offset.to_pandas()) == 1
 
 
+@pytest.mark.asyncio
+async def test_query_to_pandas_kwargs(table, table_async):
+    sync_df = (
+        LanceVectorQueryBuilder(table, [0, 0], "vector")
+        .select(["id"])
+        .limit(1)
+        .to_pandas(split_blocks=True)
+    )
+    assert sync_df["id"].tolist() == [1]
+
+    async_df = await (
+        table_async.query().select(["id"]).limit(2).to_pandas(split_blocks=True)
+    )
+    assert async_df["id"].tolist() == [1, 2]
+
+
+def test_order_by_plain_query(mem_db):
+    table = mem_db.create_table(
+        "test_order_by",
+        pa.table(
+            {
+                "group": [1, 1, 1, 2],
+                "score": [None, 1.0, 1.0, 0.5],
+                "name": ["z", "b", "a", "c"],
+            }
+        ),
+    )
+
+    res = (
+        table.search()
+        .order_by(
+            [
+                ColumnOrdering(column_name="group", ascending=True, nulls_first=False),
+                ColumnOrdering(column_name="score", ascending=True, nulls_first=True),
+                ColumnOrdering(column_name="name", ascending=True, nulls_first=False),
+            ]
+        )
+        .to_arrow()
+    )
+
+    assert res.select(["group", "score", "name"]).to_pylist() == [
+        {"group": 1, "score": None, "name": "z"},
+        {"group": 1, "score": 1.0, "name": "a"},
+        {"group": 1, "score": 1.0, "name": "b"},
+        {"group": 2, "score": 0.5, "name": "c"},
+    ]
+
+
+@pytest.mark.asyncio
+async def test_order_by_async_query(mem_db_async: AsyncConnection):
+    table = await mem_db_async.create_table(
+        "test_order_by_async",
+        pa.table(
+            {
+                "group": [1, 1, 1, 2],
+                "score": [None, 1.0, 1.0, 0.5],
+                "name": ["z", "b", "a", "c"],
+            }
+        ),
+    )
+
+    res = await (
+        table.query()
+        .order_by(
+            [
+                ColumnOrdering(column_name="group", ascending=True, nulls_first=False),
+                ColumnOrdering(column_name="score", ascending=True, nulls_first=True),
+                ColumnOrdering(column_name="name", ascending=True, nulls_first=False),
+            ]
+        )
+        .to_arrow()
+    )
+
+    assert res.select(["group", "score", "name"]).to_pylist() == [
+        {"group": 1, "score": None, "name": "z"},
+        {"group": 1, "score": 1.0, "name": "a"},
+        {"group": 1, "score": 1.0, "name": "b"},
+        {"group": 2, "score": 0.5, "name": "c"},
+    ]
+
+
 def test_query_builder(table):
     rs = (
         LanceVectorQueryBuilder(table, [0, 0], "vector")

python/python/tests/test_remote_db.py

@@ -16,6 +16,7 @@ from packaging.version import Version
 
 import lancedb
 from lancedb.conftest import MockTextEmbeddingFunction
+from lancedb.query import ColumnOrdering
 from lancedb.remote import ClientConfig
 from lancedb.remote.errors import HttpError, RetryError
 import pytest
@@ -268,6 +269,25 @@ def test_table_unimplemented_functions():
             table.to_pandas()
 
 
+def test_table_to_pandas_not_supported():
+    def handler(request):
+        if request.path == "/v1/table/test/create/?mode=create":
+            request.send_response(200)
+            request.send_header("Content-Type", "application/json")
+            request.end_headers()
+            request.wfile.write(b"{}")
+        else:
+            request.send_response(404)
+            request.end_headers()
+
+    with mock_lancedb_connection(handler) as db:
+        table = db.create_table("test", [{"id": 1}])
+        with pytest.raises(NotImplementedError):
+            table.to_pandas()
+        with pytest.raises(NotImplementedError):
+            table.to_pandas(blob_mode="bytes", split_blocks=True)
+
+
 def test_table_add_in_threadpool():
     def handler(request):
         if request.path == "/v1/table/test/insert/":
@@ -660,6 +680,18 @@ def test_query_sync_maximal():
             "ef": None,
             "filter": "id > 0",
             "columns": ["id", "name"],
+            "order_by": [
+                {
+                    "column_name": "score",
+                    "ascending": False,
+                    "nulls_first": True,
+                },
+                {
+                    "column_name": "id",
+                    "ascending": True,
+                    "nulls_first": False,
+                },
+            ],
             "vector_column": "vector2",
             "fast_search": True,
             "with_row_id": True,
@@ -677,6 +709,14 @@ def test_query_sync_maximal():
             .refine_factor(10)
             .nprobes(5)
             .where("id > 0", prefilter=True)
+            .order_by(
+                [
+                    ColumnOrdering(
+                        column_name="score", ascending=False, nulls_first=True
+                    ),
+                    ColumnOrdering(column_name="id", ascending=True, nulls_first=False),
+                ]
+            )
             .with_row_id(True)
             .select(["id", "name"])
             .to_list()

python/python/tests/test_table.py

@@ -47,6 +47,85 @@ def test_basic(mem_db: DBConnection):
     assert table.to_arrow() == expected_data
 
 
+def test_table_to_pandas_default_matches_arrow(tmp_db: DBConnection):
+    pd = pytest.importorskip("pandas")
+    data = pa.table({"id": [1, 2], "text": ["one", "two"]})
+    table = tmp_db.create_table("test_to_pandas_old_call", data=data)
+
+    expected = data.to_pandas()
+    pd.testing.assert_frame_equal(table.to_pandas(), expected)
+
+
+def test_table_to_pandas_blob_bytes(tmp_db: DBConnection):
+    pytest.importorskip("lance")
+    data = pa.table(
+        {
+            "id": pa.array([1, 2], pa.int64()),
+            "blob": pa.array([b"hello", b"world"], pa.large_binary()),
+        },
+        schema=pa.schema(
+            [
+                pa.field("id", pa.int64()),
+                pa.field(
+                    "blob", pa.large_binary(), metadata={"lance-encoding:blob": "true"}
+                ),
+            ]
+        ),
+    )
+    table = tmp_db.create_table("test_to_pandas_blob_bytes", data=data)
+
+    df = table.to_pandas(blob_mode="bytes")
+
+    assert df["blob"].tolist() == [b"hello", b"world"]
+
+
+def test_table_to_pandas_kwargs(tmp_db: DBConnection):
+    pd = pytest.importorskip("pandas")
+    data = pa.table({"id": pa.array([1, 2], pa.int64())})
+    table = tmp_db.create_table("test_to_pandas_kwargs", data=data)
+
+    df = table.to_pandas(types_mapper=pd.ArrowDtype)
+
+    assert str(df["id"].dtype) == "int64[pyarrow]"
+
+
+@pytest.mark.asyncio
+async def test_async_table_to_pandas_blob_bytes(tmp_db_async: AsyncConnection):
+    pytest.importorskip("lance")
+    data = pa.table(
+        {
+            "id": pa.array([1, 2], pa.int64()),
+            "blob": pa.array([b"hello", b"world"], pa.large_binary()),
+        },
+        schema=pa.schema(
+            [
+                pa.field("id", pa.int64()),
+                pa.field(
+                    "blob", pa.large_binary(), metadata={"lance-encoding:blob": "true"}
+                ),
+            ]
+        ),
+    )
+    table = await tmp_db_async.create_table(
+        "test_async_to_pandas_blob_bytes", data=data
+    )
+
+    df = await table.to_pandas(blob_mode="bytes")
+
+    assert df["blob"].tolist() == [b"hello", b"world"]
+
+
+@pytest.mark.asyncio
+async def test_async_table_to_pandas_kwargs(tmp_db_async: AsyncConnection):
+    pd = pytest.importorskip("pandas")
+    data = pa.table({"id": pa.array([1, 2], pa.int64())})
+    table = await tmp_db_async.create_table("test_async_to_pandas_kwargs", data=data)
+
+    df = await table.to_pandas(types_mapper=pd.ArrowDtype)
+
+    assert str(df["id"].dtype) == "int64[pyarrow]"
+
+
 def test_create_table_infers_large_int_vectors(mem_db: DBConnection):
     data = [{"vector": [0, 300]}]
 
@@ -1811,6 +1890,55 @@ def test_create_scalar_index(mem_db: DBConnection):
     assert scalar_index.name == "custom_y_index"
 
 
+def test_create_index_nested_field_paths(mem_db: DBConnection):
+    schema = pa.schema(
+        [
+            pa.field("metadata", pa.struct([pa.field("user_id", pa.int32())])),
+            pa.field(
+                "image",
+                pa.struct([pa.field("embedding", pa.list_(pa.float32(), 2))]),
+            ),
+        ]
+    )
+    data = pa.Table.from_pylist(
+        [
+            {
+                "metadata": {"user_id": i},
+                "image": {"embedding": [float(i), float(i + 1)]},
+            }
+            for i in range(256)
+        ],
+        schema=schema,
+    )
+    table = mem_db.create_table("nested_index_paths", data=data)
+
+    table.create_scalar_index("metadata.user_id", name="metadata_user_id_idx")
+    table.create_index(
+        vector_column_name="image.embedding",
+        num_partitions=1,
+        num_sub_vectors=1,
+        name="image_embedding_idx",
+    )
+
+    indices = sorted(table.list_indices(), key=lambda idx: idx.name)
+    assert [(idx.name, idx.index_type, idx.columns) for idx in indices] == [
+        ("image_embedding_idx", "IvfPq", ["image.embedding"]),
+        ("metadata_user_id_idx", "BTree", ["metadata.user_id"]),
+    ]
+
+    vector_results = (
+        table.search([0.0, 1.0], vector_column_name="image.embedding")
+        .limit(1)
+        .to_list()
+    )
+    assert len(vector_results) == 1
+    assert vector_results[0]["metadata"]["user_id"] == 0
+
+    filtered_results = table.search().where("metadata.user_id = 42").limit(1).to_list()
+    assert len(filtered_results) == 1
+    assert filtered_results[0]["metadata"]["user_id"] == 42
+
+
 def test_empty_query(mem_db: DBConnection):
     table = mem_db.create_table(
         "my_table",

python/src/connection.rs

@@ -395,12 +395,17 @@ impl Connection {
         future_into_py(py, async move {
             use lance_namespace::models::CreateNamespaceRequest;
             // 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()),
-                _ => None,
-            });
+            let mode_str = mode
+                .map(|m| match m.to_lowercase().as_str() {
+                    "create" => Ok("Create".to_string()),
+                    "exist_ok" => Ok("ExistOk".to_string()),
+                    "overwrite" => Ok("Overwrite".to_string()),
+                    _ => Err(PyValueError::new_err(format!(
+                        "Invalid mode {:?}: expected one of 'create', 'exist_ok', 'overwrite'",
+                        m
+                    ))),
+                })
+                .transpose()?;
             let request = CreateNamespaceRequest {
                 id: Some(namespace_path),
                 mode: mode_str,
@@ -428,16 +433,26 @@ impl Connection {
         future_into_py(py, async move {
             use lance_namespace::models::DropNamespaceRequest;
             // 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_str = behavior.and_then(|b| match b.to_uppercase().as_str() {
-                "RESTRICT" => Some("Restrict".to_string()),
-                "CASCADE" => Some("Cascade".to_string()),
-                _ => None,
-            });
+            let mode_str = mode
+                .map(|m| match m.to_uppercase().as_str() {
+                    "SKIP" => Ok("Skip".to_string()),
+                    "FAIL" => Ok("Fail".to_string()),
+                    _ => Err(PyValueError::new_err(format!(
+                        "Invalid mode {:?}: expected one of 'skip', 'fail'",
+                        m
+                    ))),
+                })
+                .transpose()?;
+            let behavior_str = behavior
+                .map(|b| match b.to_uppercase().as_str() {
+                    "RESTRICT" => Ok("Restrict".to_string()),
+                    "CASCADE" => Ok("Cascade".to_string()),
+                    _ => Err(PyValueError::new_err(format!(
+                        "Invalid behavior {:?}: expected one of 'restrict', 'cascade'",
+                        b
+                    ))),
+                })
+                .transpose()?;
             let request = DropNamespaceRequest {
                 id: Some(namespace_path),
                 mode: mode_str,

python/src/expr.rs

@@ -8,7 +8,9 @@
 //! DataFusion [`Expr`] nodes, bypassing SQL string parsing.
 
 use arrow::{datatypes::DataType, pyarrow::PyArrowType};
+use datafusion_common::ScalarValue;
 use lancedb::expr::{DfExpr, col as ldb_col, contains, expr_cast, lit as df_lit, lower, upper};
+use pyo3::types::PyBytes;
 use pyo3::{Bound, PyAny, PyResult, exceptions::PyValueError, prelude::*, pyfunction};
 
 /// A type-safe DataFusion expression.
@@ -141,7 +143,7 @@ pub fn expr_col(name: &str) -> PyExpr {
 
 /// Create a literal value expression.
 ///
-/// Supported Python types: `bool`, `int`, `float`, `str`.
+/// Supported Python types: `bool`, `int`, `float`, `str`, `bytes`.
 #[pyfunction]
 pub fn expr_lit(value: Bound<'_, PyAny>) -> PyResult<PyExpr> {
     // bool must be checked before int because bool is a subclass of int in Python
@@ -157,8 +159,12 @@ pub fn expr_lit(value: Bound<'_, PyAny>) -> PyResult<PyExpr> {
     if let Ok(s) = value.extract::<String>() {
         return Ok(PyExpr(df_lit(s)));
     }
+    if value.is_instance_of::<PyBytes>() {
+        let bytes = value.extract::<Vec<u8>>()?;
+        return Ok(PyExpr(df_lit(ScalarValue::Binary(Some(bytes)))));
+    }
     Err(PyValueError::new_err(format!(
-        "unsupported literal type: {}. Supported: bool, int, float, str",
+        "unsupported literal type: {}. Supported: bool, int, float, str, bytes",
         value.get_type().name()?
     )))
 }

python/src/lib.rs

@@ -15,8 +15,8 @@ use pyo3::{
 use query::{FTSQuery, HybridQuery, Query, VectorQuery};
 use session::Session;
 use table::{
-    AddColumnsResult, AddResult, AlterColumnsResult, DeleteResult, DropColumnsResult, MergeResult,
-    Table, UpdateResult,
+    AddColumnsResult, AddResult, AlterColumnsResult, DeleteResult, DropColumnsResult, LsmWriteSpec,
+    MergeResult, Table, UpdateResult,
 };
 
 pub mod arrow;
@@ -52,6 +52,7 @@ pub fn _lancedb(_py: Python, m: &Bound<'_, PyModule>) -> PyResult<()> {
     m.add_class::<AlterColumnsResult>()?;
     m.add_class::<AddResult>()?;
     m.add_class::<MergeResult>()?;
+    m.add_class::<LsmWriteSpec>()?;
     m.add_class::<DeleteResult>()?;
     m.add_class::<DropColumnsResult>()?;
     m.add_class::<UpdateResult>()?;

python/src/query.rs

@@ -23,7 +23,7 @@ use lancedb::query::QueryBase;
 use lancedb::query::QueryExecutionOptions;
 use lancedb::query::QueryFilter;
 use lancedb::query::{
-    ExecutableQuery, Query as LanceDbQuery, Select, TakeQuery as LanceDbTakeQuery,
+    ColumnOrdering, ExecutableQuery, Query as LanceDbQuery, Select, TakeQuery as LanceDbTakeQuery,
     VectorQuery as LanceDbVectorQuery,
 };
 use lancedb::table::AnyQuery;
@@ -207,6 +207,48 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
 #[derive(Clone)]
 pub struct PyQueryVectors(Vec<Arc<dyn Array>>);
 
+#[derive(Clone, FromPyObject)]
+#[pyo3(from_item_all)]
+pub struct PyColumnOrdering {
+    pub column_name: String,
+    pub ascending: bool,
+    pub nulls_first: bool,
+}
+
+impl From<ColumnOrdering> for PyColumnOrdering {
+    fn from(ordering: ColumnOrdering) -> Self {
+        Self {
+            column_name: ordering.column_name,
+            ascending: ordering.ascending,
+            nulls_first: ordering.nulls_first,
+        }
+    }
+}
+
+impl From<PyColumnOrdering> for ColumnOrdering {
+    fn from(ordering: PyColumnOrdering) -> Self {
+        Self {
+            column_name: ordering.column_name,
+            ascending: ordering.ascending,
+            nulls_first: ordering.nulls_first,
+        }
+    }
+}
+
+impl<'py> IntoPyObject<'py> for PyColumnOrdering {
+    type Target = PyDict;
+    type Output = Bound<'py, Self::Target>;
+    type Error = PyErr;
+
+    fn into_pyobject(self, py: pyo3::Python<'py>) -> PyResult<Self::Output> {
+        let dict = PyDict::new(py);
+        dict.set_item("column_name", self.column_name)?;
+        dict.set_item("ascending", self.ascending)?;
+        dict.set_item("nulls_first", self.nulls_first)?;
+        Ok(dict)
+    }
+}
+
 impl<'py> IntoPyObject<'py> for PyQueryVectors {
     type Target = PyList;
     type Output = Bound<'py, Self::Target>;
@@ -246,6 +288,7 @@ pub struct PyQueryRequest {
     pub bypass_vector_index: Option<bool>,
     pub postfilter: Option<bool>,
     pub norm: Option<String>,
+    pub order_by: Option<Vec<PyColumnOrdering>>,
 }
 
 impl From<AnyQuery> for PyQueryRequest {
@@ -273,6 +316,9 @@ impl From<AnyQuery> for PyQueryRequest {
                 bypass_vector_index: None,
                 postfilter: None,
                 norm: None,
+                order_by: query_request
+                    .order_by
+                    .map(|order_by| order_by.into_iter().map(PyColumnOrdering::from).collect()),
             },
             AnyQuery::VectorQuery(vector_query) => Self {
                 limit: vector_query.base.limit,
@@ -297,6 +343,10 @@ impl From<AnyQuery> for PyQueryRequest {
                 bypass_vector_index: Some(!vector_query.use_index),
                 postfilter: Some(!vector_query.base.prefilter),
                 norm: vector_query.base.norm.map(|n| n.to_string()),
+                order_by: vector_query
+                    .base
+                    .order_by
+                    .map(|order_by| order_by.into_iter().map(PyColumnOrdering::from).collect()),
             },
         }
     }
@@ -475,6 +525,13 @@ impl Query {
         })
     }
 
+    pub fn order_by(&mut self, ordering: Option<Vec<PyColumnOrdering>>) -> PyResult<()> {
+        let ordering =
+            ordering.map(|ordering| ordering.into_iter().map(ColumnOrdering::from).collect());
+        self.inner = self.inner.clone().order_by(ordering);
+        Ok(())
+    }
+
     #[pyo3(signature = ())]
     pub fn output_schema(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.inner.clone();
@@ -647,6 +704,13 @@ impl FTSQuery {
         self.inner = self.inner.clone().offset(offset as usize);
     }
 
+    pub fn order_by(&mut self, ordering: Option<Vec<PyColumnOrdering>>) -> PyResult<()> {
+        let ordering =
+            ordering.map(|ordering| ordering.into_iter().map(ColumnOrdering::from).collect());
+        self.inner = self.inner.clone().order_by(ordering);
+        Ok(())
+    }
+
     pub fn fast_search(&mut self) {
         self.inner = self.inner.clone().fast_search();
     }
@@ -782,6 +846,13 @@ impl VectorQuery {
         self.inner = self.inner.clone().offset(offset as usize);
     }
 
+    pub fn order_by(&mut self, ordering: Option<Vec<PyColumnOrdering>>) -> PyResult<()> {
+        let ordering =
+            ordering.map(|ordering| ordering.into_iter().map(ColumnOrdering::from).collect());
+        self.inner = self.inner.clone().order_by(ordering);
+        Ok(())
+    }
+
     pub fn fast_search(&mut self) {
         self.inner = self.inner.clone().fast_search();
     }
@@ -954,6 +1025,12 @@ impl HybridQuery {
         self.inner_fts.offset(offset);
     }
 
+    pub fn order_by(&mut self, ordering: Option<Vec<PyColumnOrdering>>) -> PyResult<()> {
+        self.inner_vec.order_by(ordering.clone())?;
+        self.inner_fts.order_by(ordering)?;
+        Ok(())
+    }
+
     pub fn fast_search(&mut self) {
         self.inner_vec.fast_search();
         self.inner_fts.fast_search();

python/src/table.rs

@@ -171,6 +171,141 @@ impl From<lancedb::table::MergeResult> for MergeResult {
     }
 }
 
+/// Specification selecting Lance's MemWAL LSM-style write path for
+/// `merge_insert`.
+///
+/// Constructed via the `bucket(...)`, `identity(...)`, or `unsharded()`
+/// classmethods, then optionally chain `with_maintained_indexes(...)` and
+/// `with_writer_config_defaults(...)`.
+#[pyclass(from_py_object)]
+#[derive(Clone, Debug)]
+pub struct LsmWriteSpec {
+    inner: lancedb::table::LsmWriteSpec,
+}
+
+#[pymethods]
+impl LsmWriteSpec {
+    /// Hash-bucket sharding by the unenforced primary key column.
+    #[staticmethod]
+    pub fn bucket(column: String, num_buckets: u32) -> Self {
+        Self {
+            inner: lancedb::table::LsmWriteSpec::bucket(column, num_buckets),
+        }
+    }
+
+    /// Identity sharding — shard by the raw value of `column`.
+    #[staticmethod]
+    pub fn identity(column: String) -> Self {
+        Self {
+            inner: lancedb::table::LsmWriteSpec::identity(column),
+        }
+    }
+
+    /// No sharding — every `merge_insert` call writes to a single
+    /// MemWAL shard.
+    #[staticmethod]
+    pub fn unsharded() -> Self {
+        Self {
+            inner: lancedb::table::LsmWriteSpec::unsharded(),
+        }
+    }
+
+    /// Replace the list of indexes the MemWAL should keep up to date as
+    /// rows are appended. Each name must reference an index that
+    /// already exists on the table at the time `set_lsm_write_spec`
+    /// is called.
+    pub fn with_maintained_indexes(&self, indexes: Vec<String>) -> Self {
+        Self {
+            inner: self.inner.clone().with_maintained_indexes(indexes),
+        }
+    }
+
+    /// Replace the default `ShardWriter` configuration recorded in the
+    /// MemWAL index, so every writer starts from the same defaults.
+    pub fn with_writer_config_defaults(&self, defaults: HashMap<String, String>) -> Self {
+        Self {
+            inner: self.inner.clone().with_writer_config_defaults(defaults),
+        }
+    }
+
+    pub fn __repr__(&self) -> String {
+        match &self.inner {
+            lancedb::table::LsmWriteSpec::Bucket {
+                column,
+                num_buckets,
+                maintained_indexes,
+                writer_config_defaults,
+            } => format!(
+                "LsmWriteSpec.bucket(column={:?}, num_buckets={}, maintained_indexes={:?}, writer_config_defaults={:?})",
+                column, num_buckets, maintained_indexes, writer_config_defaults,
+            ),
+            lancedb::table::LsmWriteSpec::Identity {
+                column,
+                maintained_indexes,
+                writer_config_defaults,
+            } => format!(
+                "LsmWriteSpec.identity(column={:?}, maintained_indexes={:?}, writer_config_defaults={:?})",
+                column, maintained_indexes, writer_config_defaults,
+            ),
+            lancedb::table::LsmWriteSpec::Unsharded {
+                maintained_indexes,
+                writer_config_defaults,
+            } => format!(
+                "LsmWriteSpec.unsharded(maintained_indexes={:?}, writer_config_defaults={:?})",
+                maintained_indexes, writer_config_defaults,
+            ),
+        }
+    }
+
+    /// Discriminator string identifying the variant ("bucket", "identity",
+    /// or "unsharded").
+    #[getter]
+    pub fn spec_type(&self) -> &'static str {
+        match &self.inner {
+            lancedb::table::LsmWriteSpec::Bucket { .. } => "bucket",
+            lancedb::table::LsmWriteSpec::Identity { .. } => "identity",
+            lancedb::table::LsmWriteSpec::Unsharded { .. } => "unsharded",
+        }
+    }
+
+    /// Bucket and identity variants: the sharding column. `None` for unsharded.
+    #[getter]
+    pub fn column(&self) -> Option<String> {
+        match &self.inner {
+            lancedb::table::LsmWriteSpec::Bucket { column, .. }
+            | lancedb::table::LsmWriteSpec::Identity { column, .. } => Some(column.clone()),
+            lancedb::table::LsmWriteSpec::Unsharded { .. } => None,
+        }
+    }
+
+    /// Bucket variant only: the number of buckets.
+    #[getter]
+    pub fn num_buckets(&self) -> Option<u32> {
+        match &self.inner {
+            lancedb::table::LsmWriteSpec::Bucket { num_buckets, .. } => Some(*num_buckets),
+            _ => None,
+        }
+    }
+
+    /// Names of indexes the MemWAL should keep up to date during writes.
+    #[getter]
+    pub fn maintained_indexes(&self) -> Vec<String> {
+        self.inner.maintained_indexes().to_vec()
+    }
+
+    /// Default `ShardWriter` configuration recorded by this spec.
+    #[getter]
+    pub fn writer_config_defaults(&self) -> HashMap<String, String> {
+        self.inner.writer_config_defaults().clone()
+    }
+}
+
+impl From<LsmWriteSpec> for lancedb::table::LsmWriteSpec {
+    fn from(spec: LsmWriteSpec) -> Self {
+        spec.inner
+    }
+}
+
 #[pyclass(get_all, from_py_object)]
 #[derive(Clone, Debug)]
 pub struct AddColumnsResult {
@@ -805,6 +940,37 @@ impl Table {
         })
     }
 
+    pub fn set_unenforced_primary_key<'a>(
+        self_: PyRef<'a, Self>,
+        columns: Vec<String>,
+    ) -> PyResult<Bound<'a, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        future_into_py(self_.py(), async move {
+            inner
+                .set_unenforced_primary_key(columns)
+                .await
+                .infer_error()
+        })
+    }
+
+    pub fn set_lsm_write_spec<'a>(
+        self_: PyRef<'a, Self>,
+        spec: LsmWriteSpec,
+    ) -> PyResult<Bound<'a, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        let native_spec = lancedb::table::LsmWriteSpec::from(spec);
+        future_into_py(self_.py(), async move {
+            inner.set_lsm_write_spec(native_spec).await.infer_error()
+        })
+    }
+
+    pub fn unset_lsm_write_spec(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
+        let inner = self_.inner_ref()?.clone();
+        future_into_py(self_.py(), async move {
+            inner.unset_lsm_write_spec().await.infer_error()
+        })
+    }
+
     pub fn uses_v2_manifest_paths(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {

python/tests/test_expr.py

@@ -33,6 +33,14 @@ class TestExprConstruction:
         e = lit(True)
         assert isinstance(e, Expr)
 
+    def test_lit_bytes(self):
+        e = lit(b"\xde\xad\xbe\xef")
+        assert isinstance(e, Expr)
+
+    def test_lit_bytes_empty(self):
+        e = lit(b"")
+        assert isinstance(e, Expr)
+
     def test_lit_unsupported_type_raises(self):
         with pytest.raises(Exception):
             lit([1, 2, 3])
@@ -135,6 +143,43 @@ class TestExprOperators:
         assert e.to_sql() == "(name = 'alice')"
 
 
+class TestExprBytesLiteral:
+    def test_bytes_to_sql(self):
+        e = lit(b"\xde\xad\xbe\xef")
+        assert e.to_sql() == "X'DEADBEEF'"
+
+    def test_empty_bytes_to_sql(self):
+        e = lit(b"")
+        assert e.to_sql() == "X''"
+
+    def test_bytes_repr(self):
+        e = lit(b"\x01\x02")
+        assert repr(e) == "Expr(X'0102')"
+
+    def test_bytes_equality_expr_sql(self):
+        e = col("data") == lit(b"\xca\xfe")
+        assert e.to_sql() == "(data = X'CAFE')"
+
+    def test_bytes_ne_expr_sql(self):
+        e = col("data") != lit(b"\xff")
+        assert e.to_sql() == "(data <> X'FF')"
+
+    def test_bytes_compound_expr_sql(self):
+        e = (col("data") == lit(b"\x01")) & (col("id") > lit(5))
+        assert e.to_sql() == "((data = X'01') AND (id > 5))"
+
+    def test_bytes_in_function_call(self):
+        # Regression test: binary literals inside scalar function calls
+        # used to fail because DataFusion's unparser does not support Binary
+        # scalars.  Now handled via a placeholder-substitution rewrite.
+        e = func("contains", col("data"), lit(b"\xff"))
+        assert e.to_sql() == "contains(data, X'FF')"
+
+    def test_bytes_in_not(self):
+        e = ~(col("data") == lit(b"\xff"))
+        assert e.to_sql() == "NOT (data = X'FF')"
+
+
 class TestExprStringMethods:
     def test_lower(self):
         e = col("name").lower()
@@ -385,3 +430,44 @@ class TestColNamingIntegration:
         )
         assert "upper_name" in result.schema.names
         assert sorted(result["upper_name"].to_pylist()) == ["ALICE", "BOB", "CHARLIE"]
+
+
+# ── bytes / binary column integration tests ───────────────────────────────────
+
+
+@pytest.fixture
+def binary_table(tmp_path):
+    db = lancedb.connect(str(tmp_path))
+    data = pa.table(
+        {
+            "id": [1, 2, 3],
+            "payload": pa.array(
+                [b"\x01\x02", b"\xca\xfe", b"\xff\x00"],
+                type=pa.binary(),
+            ),
+        }
+    )
+    return db.create_table("binary_test", data)
+
+
+class TestExprBytesIntegration:
+    def test_binary_equality_filter(self, binary_table):
+        result = (
+            binary_table.search().where(col("payload") == lit(b"\xca\xfe")).to_arrow()
+        )
+        assert result.num_rows == 1
+        assert result["id"][0].as_py() == 2
+
+    def test_binary_ne_filter(self, binary_table):
+        result = (
+            binary_table.search().where(col("payload") != lit(b"\x01\x02")).to_arrow()
+        )
+        assert result.num_rows == 2
+
+    def test_binary_compound_filter(self, binary_table):
+        result = (
+            binary_table.search()
+            .where((col("payload") == lit(b"\x01\x02")) | (col("id") == lit(3)))
+            .to_arrow()
+        )
+        assert result.num_rows == 2

rust/lancedb/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.28.0-beta.11"
+version = "0.29.1-beta.0"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true