Bump version: 0.24.0-beta.0 → 0.24.0

lance [release
details](https://github.com/lancedb/lance/releases/tag/v0.30.0)
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Chores**
- Updated dependency specifications to use exact version numbers instead
of referencing a git repository and tag.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Signed-off-by: BubbleCal <bubble-cal@outlook.com>

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.19.0-beta.6"
+current_version = "0.20.1-beta.2"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/workflows/docs.yml

@@ -18,17 +18,24 @@ concurrency:
   group: "pages"
   cancel-in-progress: true
 
+env:
+  # This reduces the disk space needed for the build
+  RUSTFLAGS: "-C debuginfo=0"
+  # according to: https://matklad.github.io/2021/09/04/fast-rust-builds.html
+  # CI builds are faster with incremental disabled.
+  CARGO_INCREMENTAL: "0"
+
 jobs:
   # Single deploy job since we're just deploying
   build:
     environment:
       name: github-pages
       url: ${{ steps.deployment.outputs.page_url }}
-    runs-on: buildjet-8vcpu-ubuntu-2204
+    runs-on: ubuntu-24.04
     steps:
       - name: Checkout
         uses: actions/checkout@v4
-      - name: Install dependecies needed for ubuntu
+      - name: Install dependencies needed for ubuntu
         run: |
           sudo apt install -y protobuf-compiler libssl-dev
           rustup update && rustup default
@@ -38,6 +45,7 @@ jobs:
           python-version: "3.10"
           cache: "pip"
           cache-dependency-path: "docs/requirements.txt"
+      - uses: Swatinem/rust-cache@v2
       - name: Build Python
         working-directory: python
         run: |
@@ -49,7 +57,6 @@ jobs:
           node-version: 20
           cache: 'npm'
           cache-dependency-path: node/package-lock.json
-      - uses: Swatinem/rust-cache@v2
       - name: Install node dependencies
         working-directory: node
         run: |

.github/workflows/java.yml

@@ -35,6 +35,9 @@ jobs:
       - uses: Swatinem/rust-cache@v2
         with:
           workspaces: java/core/lancedb-jni
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
+        with:
+          components: rustfmt
       - name: Run cargo fmt
         run: cargo fmt --check
         working-directory: ./java/core/lancedb-jni
@@ -68,6 +71,9 @@ jobs:
       - uses: Swatinem/rust-cache@v2
         with:
           workspaces: java/core/lancedb-jni
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
+        with:
+          components: rustfmt
       - name: Run cargo fmt
         run: cargo fmt --check
         working-directory: ./java/core/lancedb-jni
@@ -110,4 +116,3 @@ jobs:
           -Djdk.reflect.useDirectMethodHandle=false \
           -Dio.netty.tryReflectionSetAccessible=true"
           JAVA_HOME=$JAVA_17 mvn clean test
-  

.github/workflows/make-release-commit.yml

@@ -84,6 +84,7 @@ jobs:
         run: |
           pip install bump-my-version PyGithub packaging
           bash ci/bump_version.sh ${{ inputs.type }} ${{ inputs.bump-minor }} v $COMMIT_BEFORE_BUMP
+          bash ci/update_lockfiles.sh --amend
       - name: Push new version tag
         if: ${{ !inputs.dry_run }}
         uses: ad-m/github-push-action@master
@@ -92,11 +93,3 @@ jobs:
           github_token: ${{ secrets.LANCEDB_RELEASE_TOKEN }}
           branch: ${{ github.ref }}
           tags: true
-      - uses: ./.github/workflows/update_package_lock
-        if: ${{ !inputs.dry_run && inputs.other }}
-        with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
-      - uses: ./.github/workflows/update_package_lock_nodejs
-        if: ${{ !inputs.dry_run && inputs.other }}
-        with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/nodejs.yml

@@ -47,6 +47,9 @@ jobs:
       run: |
         sudo apt update
         sudo apt install -y protobuf-compiler libssl-dev
+    - uses: actions-rust-lang/setup-rust-toolchain@v1
+      with:
+        components: rustfmt, clippy
     - name: Lint
       run: |
         cargo fmt --all -- --check
@@ -113,7 +116,7 @@ jobs:
         set -e
         npm ci
         npm run docs
-        if ! git diff --exit-code; then
+        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

.github/workflows/npm-publish.yml

@@ -505,6 +505,8 @@ jobs:
     name: vectordb NPM Publish
     needs: [node, node-macos, node-linux-gnu, node-windows]
     runs-on: ubuntu-latest
+    permissions:
+      contents: write
     # Only runs on tags that matches the make-release action
     if: startsWith(github.ref, 'refs/tags/v')
     steps:
@@ -537,6 +539,20 @@ jobs:
         # We need to deprecate the old package to avoid confusion.
         # Each time we publish a new version, it gets undeprecated.
         run: npm deprecate vectordb "Use @lancedb/lancedb instead."
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          ref: main
+      - name: Update package-lock.json
+        run: |
+          git config user.name 'Lance Release'
+          git config user.email 'lance-dev@lancedb.com'
+          bash ci/update_lockfiles.sh
+      - name: Push new commit
+        uses: ad-m/github-push-action@master
+        with:
+          github_token: ${{ secrets.LANCEDB_RELEASE_TOKEN }}
+          branch: main
       - name: Notify Slack Action
         uses: ravsamhq/notify-slack-action@2.3.0
         if: ${{ always() }}
@@ -546,21 +562,3 @@ jobs:
           notification_title: "{workflow} is failing"
         env:
           SLACK_WEBHOOK_URL: ${{ secrets.ACTION_MONITORING_SLACK }}
-
-  update-package-lock:
-    if: startsWith(github.ref, 'refs/tags/v')
-    needs: [release]
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-        with:
-          ref: main
-          token: ${{ secrets.LANCEDB_RELEASE_TOKEN }}
-          fetch-depth: 0
-          lfs: true
-      - uses: ./.github/workflows/update_package_lock
-        with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/python.yml

@@ -136,9 +136,9 @@ jobs:
       - uses: ./.github/workflows/run_tests
         with:
           integration: true
-      - name: Test without pylance
+      - name: Test without pylance or pandas
         run: |
-          pip uninstall -y pylance
+          pip uninstall -y pylance pandas
           pytest -vv python/tests/test_table.py
       # Make sure wheels are not included in the Rust cache
       - name: Delete wheels
@@ -228,6 +228,7 @@ jobs:
       - name: Install lancedb
         run: |
           pip install "pydantic<2"
+          pip install pyarrow==16
           pip install --extra-index-url https://pypi.fury.io/lancedb/ -e .[tests]
           pip install tantivy
       - name: Run tests

.github/workflows/run_tests/action.yml

@@ -24,8 +24,8 @@ runs:
     - name: pytest (with integration)
       shell: bash
       if: ${{ inputs.integration == 'true' }}
-      run: pytest -m "not slow" -x -v --durations=30 python/python/tests
+      run: pytest -m "not slow" -vv --durations=30 python/python/tests
     - name: pytest (no integration tests)
       shell: bash
       if: ${{ inputs.integration != 'true' }}
-      run: pytest -m "not slow and not s3_test" -x -v --durations=30 python/python/tests
+      run: pytest -m "not slow and not s3_test" -vv --durations=30 python/python/tests

.github/workflows/rust.yml

@@ -40,6 +40,9 @@ jobs:
         with:
           fetch-depth: 0
           lfs: true
+      - uses: actions-rust-lang/setup-rust-toolchain@v1
+        with:
+          components: rustfmt, clippy
       - uses: Swatinem/rust-cache@v2
         with:
           workspaces: rust
@@ -160,8 +163,8 @@ jobs:
     strategy:
       matrix:
         target:
-        - x86_64-pc-windows-msvc
+          - x86_64-pc-windows-msvc
-        - aarch64-pc-windows-msvc
+          - aarch64-pc-windows-msvc
     defaults:
       run:
         working-directory: rust/lancedb

.github/workflows/update_package_lock/action.yml

@@ -1,33 +0,0 @@
-name: update_package_lock
-description: "Update node's package.lock"
-
-inputs:
-  github_token:
-    required: true
-    description: "github token for the repo"
-
-runs:
-  using: "composite"
-  steps:
-    - uses: actions/setup-node@v3
-      with:
-        node-version: 20
-    - name: Set git configs
-      shell: bash
-      run: |
-        git config user.name 'Lance Release'
-        git config user.email 'lance-dev@lancedb.com'
-    - name: Update package-lock.json file
-      working-directory: ./node
-      run: |
-        npm install
-        git add package-lock.json
-        git commit -m "Updating package-lock.json"
-      shell: bash
-    - name: Push changes
-      if: ${{ inputs.dry_run }} == "false"
-      uses: ad-m/github-push-action@master
-      with:
-        github_token: ${{ inputs.github_token }}
-        branch: main
-        tags: true

.github/workflows/update_package_lock_nodejs/action.yml

@@ -1,33 +0,0 @@
-name: update_package_lock_nodejs
-description: "Update nodejs's package.lock"
-
-inputs:
-  github_token:
-    required: true
-    description: "github token for the repo"
-
-runs:
-  using: "composite"
-  steps:
-    - uses: actions/setup-node@v3
-      with:
-        node-version: 20
-    - name: Set git configs
-      shell: bash
-      run: |
-        git config user.name 'Lance Release'
-        git config user.email 'lance-dev@lancedb.com'
-    - name: Update package-lock.json file
-      working-directory: ./nodejs
-      run: |
-        npm install
-        git add package-lock.json
-        git commit -m "Updating package-lock.json"
-      shell: bash
-    - name: Push changes
-      if: ${{ inputs.dry_run }} == "false"
-      uses: ad-m/github-push-action@master
-      with:
-        github_token: ${{ inputs.github_token }}
-        branch: main
-        tags: true

Cargo.toml

@@ -21,34 +21,32 @@ categories = ["database-implementations"]
 rust-version = "1.78.0"
 
 [workspace.dependencies]
-lance = { "version" = "=0.25.3", "features" = [
+lance = { "version" = "=0.30.0", "features" = ["dynamodb"] }
-    "dynamodb",
+lance-io = "=0.30.0"
-], tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-index = "=0.30.0"
-lance-io = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-linalg = "=0.30.0"
-lance-index = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-table = "=0.30.0"
-lance-linalg = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-testing = "=0.30.0"
-lance-table = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-datafusion = "=0.30.0"
-lance-testing = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
+lance-encoding = "=0.30.0"
-lance-datafusion = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
-lance-encoding = { version = "=0.25.3", tag = "v0.25.3-beta.4", git = "https://github.com/lancedb/lance" }
 # Note that this one does not include pyarrow
-arrow = { version = "54.1", optional = false }
+arrow = { version = "55.1", optional = false }
-arrow-array = "54.1"
+arrow-array = "55.1"
-arrow-data = "54.1"
+arrow-data = "55.1"
-arrow-ipc = "54.1"
+arrow-ipc = "55.1"
-arrow-ord = "54.1"
+arrow-ord = "55.1"
-arrow-schema = "54.1"
+arrow-schema = "55.1"
-arrow-arith = "54.1"
+arrow-arith = "55.1"
-arrow-cast = "54.1"
+arrow-cast = "55.1"
 async-trait = "0"
-datafusion = { version = "46.0", default-features = false }
+datafusion = { version = "47.0", default-features = false }
-datafusion-catalog = "46.0"
+datafusion-catalog = "47.0"
-datafusion-common = { version = "46.0", default-features = false }
+datafusion-common = { version = "47.0", default-features = false }
-datafusion-execution = "46.0"
+datafusion-execution = "47.0"
-datafusion-expr = "46.0"
+datafusion-expr = "47.0"
-datafusion-physical-plan = "46.0"
+datafusion-physical-plan = "47.0"
 env_logger = "0.11"
-half = { "version" = "=2.4.1", default-features = false, features = [
+half = { "version" = "=2.5.0", default-features = false, features = [
     "num-traits",
 ] }
 futures = "0"
@@ -59,19 +57,16 @@ pin-project = "1.0.7"
 snafu = "0.8"
 url = "2"
 num-traits = "0.2"
-rand = "0.8"
+rand = "0.9"
 regex = "1.10"
 lazy_static = "1"
 semver = "1.0.25"
-
 # Temporary pins to work around downstream issues
 # https://github.com/apache/arrow-rs/commit/2fddf85afcd20110ce783ed5b4cdeb82293da30b
-chrono = "=0.4.39"
+chrono = "=0.4.41"
 # https://github.com/RustCrypto/formats/issues/1684
 base64ct = "=1.6.0"
-
 # Workaround for: https://github.com/eira-fransham/crunchy/issues/13
 crunchy = "=0.2.2"
-
 # Workaround for: https://github.com/Lokathor/bytemuck/issues/306
 bytemuck_derive = ">=1.8.1, <1.9.0"

README.md

@@ -1,94 +1,97 @@
 <a href="https://cloud.lancedb.com" target="_blank">
   <img src="https://github.com/user-attachments/assets/92dad0a2-2a37-4ce1-b783-0d1b4f30a00c" alt="LanceDB Cloud Public Beta" width="100%" style="max-width: 100%;">
 </a>
-
 <div align="center">
-<p align="center">
 
-<picture>
+[![LanceDB](docs/src/assets/hero-header.png)](https://lancedb.com)
-  <source media="(prefers-color-scheme: dark)" srcset="https://github.com/user-attachments/assets/ac270358-333e-4bea-a132-acefaa94040e">
+[![Website](https://img.shields.io/badge/-Website-100000?style=for-the-badge&labelColor=645cfb&color=645cfb)](https://lancedb.com/)
-  <source media="(prefers-color-scheme: light)" srcset="https://github.com/user-attachments/assets/b864d814-0d29-4784-8fd9-807297c758c0">
+[![Blog](https://img.shields.io/badge/Blog-100000?style=for-the-badge&labelColor=645cfb&color=645cfb)](https://blog.lancedb.com/)
-  <img alt="LanceDB Logo" src="https://github.com/user-attachments/assets/b864d814-0d29-4784-8fd9-807297c758c0" width=300>
+[![Discord](https://img.shields.io/badge/-Discord-100000?style=for-the-badge&logo=discord&logoColor=white&labelColor=645cfb&color=645cfb)](https://discord.gg/zMM32dvNtd)
-</picture>
+[![Twitter](https://img.shields.io/badge/-Twitter-100000?style=for-the-badge&logo=x&logoColor=white&labelColor=645cfb&color=645cfb)](https://twitter.com/lancedb)
+[![LinkedIn](https://img.shields.io/badge/-LinkedIn-100000?style=for-the-badge&logo=linkedin&logoColor=white&labelColor=645cfb&color=645cfb)](https://www.linkedin.com/company/lancedb/)
 
-**Search More, Manage Less**
 
-<a href='https://github.com/lancedb/vectordb-recipes/tree/main' target="_blank"><img alt='LanceDB' src='https://img.shields.io/badge/VectorDB_Recipes-100000?style=for-the-badge&logo=LanceDB&logoColor=white&labelColor=645cfb&color=645cfb'/></a>
+<img src="docs/src/assets/lancedb.png" alt="LanceDB" width="50%">
-<a href='https://lancedb.github.io/lancedb/' target="_blank"><img alt='lancdb' src='https://img.shields.io/badge/DOCS-100000?style=for-the-badge&logo=lancdb&logoColor=white&labelColor=645cfb&color=645cfb'/></a>
-[![Blog](https://img.shields.io/badge/Blog-12100E?style=for-the-badge&logoColor=white)](https://blog.lancedb.com/)
-[![Discord](https://img.shields.io/badge/Discord-%235865F2.svg?style=for-the-badge&logo=discord&logoColor=white)](https://discord.gg/zMM32dvNtd)
-[![Twitter](https://img.shields.io/badge/Twitter-%231DA1F2.svg?style=for-the-badge&logo=Twitter&logoColor=white)](https://twitter.com/lancedb)
-[![Gurubase](https://img.shields.io/badge/Gurubase-Ask%20LanceDB%20Guru-006BFF?style=for-the-badge)](https://gurubase.io/g/lancedb)
 
-</p>
+# **The Multimodal AI Lakehouse**
 
-<img max-width="750px" alt="LanceDB Multimodal Search" src="https://github.com/lancedb/lancedb/assets/917119/09c5afc5-7816-4687-bae4-f2ca194426ec">
+[**How to Install** ](#how-to-install) ✦ [**Detailed Documentation**](https://lancedb.github.io/lancedb/) ✦ [**Tutorials and Recipes**](https://github.com/lancedb/vectordb-recipes/tree/main) ✦  [**Contributors**](#contributors) 
+
+**The ultimate multimodal data platform for AI/ML applications.** 
+
+LanceDB is designed for fast, scalable, and production-ready vector search. It is built on top of the Lance columnar format. You can store, index, and search over petabytes of multimodal data and vectors with ease. 
+LanceDB is a central location where developers can build, train and analyze their AI workloads.
 
-</p>
 </div>
 
-<hr />
+<br>
 
-LanceDB is an open-source database for vector-search built with persistent storage, which greatly simplifies retrieval, filtering and management of embeddings.
+## **Demo: Multimodal Search by Keyword, Vector or with SQL**
+<img max-width="750px" alt="LanceDB Multimodal Search" src="https://github.com/lancedb/lancedb/assets/917119/09c5afc5-7816-4687-bae4-f2ca194426ec">
 
-The key features of LanceDB include:
+## **Star LanceDB to get updates!**
 
-* Production-scale vector search with no servers to manage.
+<details>
+<summary>⭐ Click here ⭐  to see how fast we're growing!</summary>
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=lancedb/lancedb&theme=dark&type=Date">
+  <img width="100%" src="https://api.star-history.com/svg?repos=lancedb/lancedb&theme=dark&type=Date">
+</picture>
+</details>
 
-* Store, query and filter vectors, metadata and multi-modal data (text, images, videos, point clouds, and more).
+## **Key Features**:
 
-* Support for vector similarity search, full-text search and SQL.
+- **Fast Vector Search**: Search billions of vectors in milliseconds with state-of-the-art indexing.
+- **Comprehensive Search**: Support for vector similarity search, full-text search and SQL.
+- **Multimodal Support**: Store, query and filter vectors, metadata and multimodal data (text, images, videos, point clouds, and more).
+- **Advanced Features**: Zero-copy, automatic versioning, manage versions of your data without needing extra infrastructure. GPU support in building vector index.
 
-* Native Python and Javascript/Typescript support.
+### **Products**:
+- **Open Source & Local**: 100% open source, runs locally or in your cloud. No vendor lock-in.
+- **Cloud and Enterprise**: Production-scale vector search with no servers to manage. Complete data sovereignty and security.
 
-* Zero-copy, automatic versioning, manage versions of your data without needing extra infrastructure.
+### **Ecosystem**:
+- **Columnar Storage**: Built on the Lance columnar format for efficient storage and analytics.
+- **Seamless Integration**: Python, Node.js, Rust, and REST APIs for easy integration. Native Python and Javascript/Typescript support.
+- **Rich Ecosystem**: Integrations with [**LangChain** 🦜️🔗](https://python.langchain.com/docs/integrations/vectorstores/lancedb/), [**LlamaIndex** 🦙](https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/LanceDBIndexDemo.html), Apache-Arrow, Pandas, Polars, DuckDB and more on the way.
 
-* GPU support in building vector index(*).
+## **How to Install**:
 
-* Ecosystem integrations with [LangChain 🦜️🔗](https://python.langchain.com/docs/integrations/vectorstores/lancedb/), [LlamaIndex 🦙](https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/LanceDBIndexDemo.html), Apache-Arrow, Pandas, Polars, DuckDB and more on the way.
+Follow the [Quickstart](https://lancedb.github.io/lancedb/basic/) doc to set up LanceDB locally. 
 
-LanceDB's core is written in Rust 🦀 and is built using <a href="https://github.com/lancedb/lance">Lance</a>, an open-source columnar format designed for performant ML workloads.
+**API & SDK:** We also support Python, Typescript and Rust SDKs
 
-## Quick Start
+| Interface | Documentation |
+|-----------|---------------|
+| Python SDK | https://lancedb.github.io/lancedb/python/python/ |
+| Typescript SDK | https://lancedb.github.io/lancedb/js/globals/ |
+| Rust SDK | https://docs.rs/lancedb/latest/lancedb/index.html |
+| REST API | https://docs.lancedb.com/api-reference/introduction |
 
-**Javascript**
+## **Join Us and Contribute**
-```shell
-npm install @lancedb/lancedb
-```
 
-```javascript
+We welcome contributions from everyone! Whether you're a developer, researcher, or just someone who wants to help out. 
-import * as lancedb from "@lancedb/lancedb";
 
-const db = await lancedb.connect("data/sample-lancedb");
+If you have any suggestions or feature requests, please feel free to open an issue on GitHub or discuss it on our [**Discord**](https://discord.gg/G5DcmnZWKB) server.
-const table = await db.createTable("vectors", [
+
-	{ id: 1, vector: [0.1, 0.2], item: "foo", price: 10 },
+[**Check out the GitHub Issues**](https://github.com/lancedb/lancedb/issues) if you would like to work on the features that are planned for the future. If you have any suggestions or feature requests, please feel free to open an issue on GitHub. 
-	{ id: 2, vector: [1.1, 1.2], item: "bar", price: 50 },
+
-], {mode: 'overwrite'});
+## **Contributors**
+
+<a href="https://github.com/lancedb/lancedb/graphs/contributors">
+  <img src="https://contrib.rocks/image?repo=lancedb/lancedb" />
+</a>
 
 
-const query = table.vectorSearch([0.1, 0.3]).limit(2);
+## **Stay in Touch With Us**
-const results = await query.toArray();
+<div align="center">
 
-// You can also search for rows by specific criteria without involving a vector search.
+</br>
-const rowsByCriteria = await table.query().where("price >= 10").toArray();
-```
 
-**Python**
+[![Website](https://img.shields.io/badge/-Website-100000?style=for-the-badge&labelColor=645cfb&color=645cfb)](https://lancedb.com/)
-```shell
+[![Blog](https://img.shields.io/badge/Blog-100000?style=for-the-badge&labelColor=645cfb&color=645cfb)](https://blog.lancedb.com/)
-pip install lancedb
+[![Discord](https://img.shields.io/badge/-Discord-100000?style=for-the-badge&logo=discord&logoColor=white&labelColor=645cfb&color=645cfb)](https://discord.gg/zMM32dvNtd)
-```
+[![Twitter](https://img.shields.io/badge/-Twitter-100000?style=for-the-badge&logo=x&logoColor=white&labelColor=645cfb&color=645cfb)](https://twitter.com/lancedb)
+[![LinkedIn](https://img.shields.io/badge/-LinkedIn-100000?style=for-the-badge&logo=linkedin&logoColor=white&labelColor=645cfb&color=645cfb)](https://www.linkedin.com/company/lancedb/)
 
-```python
+</div>
-import lancedb
-
-uri = "data/sample-lancedb"
-db = lancedb.connect(uri)
-table = db.create_table("my_table",
-                         data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
-                               {"vector": [5.9, 26.5], "item": "bar", "price": 20.0}])
-result = table.search([100, 100]).limit(2).to_pandas()
-```
-
-## Blogs, Tutorials & Videos
-* 📈 <a href="https://blog.lancedb.com/benchmarking-random-access-in-lance/">2000x better performance with Lance over Parquet</a>
-* 🤖 <a href="https://github.com/lancedb/vectordb-recipes/tree/main/examples/Youtube-Search-QA-Bot">Build a question and answer bot with LanceDB</a>

ci/set_lance_version.py

@@ -0,0 +1,174 @@
+import argparse
+import sys
+import json
+
+
+def run_command(command: str) -> str:
+    """
+    Run a shell command and return stdout as a string.
+    If exit code is not 0, raise an exception with the stderr output.
+    """
+    import subprocess
+
+    result = subprocess.run(command, shell=True, capture_output=True, text=True)
+    if result.returncode != 0:
+        raise Exception(f"Command failed with error: {result.stderr.strip()}")
+    return result.stdout.strip()
+
+
+def get_latest_stable_version() -> str:
+    version_line = run_command("cargo info lance | grep '^version:'")
+    version = version_line.split(" ")[1].strip()
+    return version
+
+
+def get_latest_preview_version() -> str:
+    lance_tags = run_command(
+        "git ls-remote --tags https://github.com/lancedb/lance.git | grep 'refs/tags/v[0-9beta.-]\\+$'"
+    ).splitlines()
+    lance_tags = (
+        tag.split("refs/tags/")[1]
+        for tag in lance_tags
+        if "refs/tags/" in tag and "beta" in tag
+    )
+    from packaging.version import Version
+
+    latest = max(
+        (tag[1:] for tag in lance_tags if tag.startswith("v")), key=lambda t: Version(t)
+    )
+    return str(latest)
+
+
+def extract_features(line: str) -> list:
+    """
+    Extracts the features from a line in Cargo.toml.
+    Example: 'lance = { "version" = "=0.29.0", "features" = ["dynamodb"] }'
+    Returns: ['dynamodb']
+    """
+    import re
+
+    match = re.search(r'"features"\s*=\s*\[(.*?)\]', line)
+    if match:
+        features_str = match.group(1)
+        return [f.strip('"') for f in features_str.split(",")]
+    return []
+
+
+def update_cargo_toml(line_updater):
+    """
+    Updates the Cargo.toml file by applying the line_updater function to each line.
+    The line_updater function should take a line as input and return the updated line.
+    """
+    with open("Cargo.toml", "r") as f:
+        lines = f.readlines()
+
+    new_lines = []
+    for line in lines:
+        if line.startswith("lance"):
+            # Update the line using the provided function
+            new_lines.append(line_updater(line))
+        else:
+            # Keep the line unchanged
+            new_lines.append(line)
+
+    with open("Cargo.toml", "w") as f:
+        f.writelines(new_lines)
+
+
+def set_stable_version(version: str):
+    """
+    Sets lines to
+    lance = { "version" = "=0.29.0", "features" = ["dynamodb"] }
+    lance-io = "=0.29.0"
+    ...
+    """
+
+    def line_updater(line: str) -> str:
+        package_name = line.split("=", maxsplit=1)[0].strip()
+        features = extract_features(line)
+        if features:
+            return f'{package_name} = {{ "version" = "={version}", "features" = {json.dumps(features)} }}\n'
+        else:
+            return f'{package_name} = "={version}"\n'
+
+    update_cargo_toml(line_updater)
+
+
+def set_preview_version(version: str):
+    """
+    Sets lines to
+    lance = { "version" = "=0.29.0", "features" = ["dynamodb"], tag = "v0.29.0-beta.2", git="https://github.com/lancedb/lance.git" }
+    lance-io = { version = "=0.29.0", tag = "v0.29.0-beta.2", git="https://github.com/lancedb/lance.git" }
+    ...
+    """
+
+    def line_updater(line: str) -> str:
+        package_name = line.split("=", maxsplit=1)[0].strip()
+        features = extract_features(line)
+        base_version = version.split("-")[0]  # Get the base version without beta suffix
+        if features:
+            return f'{package_name} = {{ "version" = "={base_version}", "features" = {json.dumps(features)}, "tag" = "v{version}", "git" = "https://github.com/lancedb/lance.git" }}\n'
+        else:
+            return f'{package_name} = {{ "version" = "={base_version}", "tag" = "v{version}", "git" = "https://github.com/lancedb/lance.git" }}\n'
+
+    update_cargo_toml(line_updater)
+
+
+def set_local_version():
+    """
+    Sets lines to
+    lance = { path = "../lance/rust/lance", features = ["dynamodb"] }
+    lance-io = { path = "../lance/rust/lance-io" }
+    ...
+    """
+
+    def line_updater(line: str) -> str:
+        package_name = line.split("=", maxsplit=1)[0].strip()
+        features = extract_features(line)
+        if features:
+            return f'{package_name} = {{ "path" = "../lance/rust/{package_name}", "features" = {json.dumps(features)} }}\n'
+        else:
+            return f'{package_name} = {{ "path" = "../lance/rust/{package_name}" }}\n'
+
+    update_cargo_toml(line_updater)
+
+
+parser = argparse.ArgumentParser(description="Set the version of the Lance package.")
+parser.add_argument(
+    "version",
+    type=str,
+    help="The version to set for the Lance package. Use 'stable' for the latest stable version, 'preview' for latest preview version, or a specific version number (e.g., '0.1.0'). You can also specify 'local' to use a local path.",
+)
+args = parser.parse_args()
+
+if args.version == "stable":
+    latest_stable_version = get_latest_stable_version()
+    print(
+        f"Found latest stable version: \033[1mv{latest_stable_version}\033[0m",
+        file=sys.stderr,
+    )
+    set_stable_version(latest_stable_version)
+elif args.version == "preview":
+    latest_preview_version = get_latest_preview_version()
+    print(
+        f"Found latest preview version: \033[1mv{latest_preview_version}\033[0m",
+        file=sys.stderr,
+    )
+    set_preview_version(latest_preview_version)
+elif args.version == "local":
+    set_local_version()
+else:
+    # Parse the version number.
+    version = args.version
+    # Ignore initial v if present.
+    if version.startswith("v"):
+        version = version[1:]
+
+    if "beta" in version:
+        set_preview_version(version)
+    else:
+        set_stable_version(version)
+
+print("Updating lockfiles...", file=sys.stderr, end="")
+run_command("cargo metadata > /dev/null")
+print(" done.", file=sys.stderr)

ci/update_lockfiles.sh

@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+AMEND=false
+
+for arg in "$@"; do
+  if [[ "$arg" == "--amend" ]]; then
+    AMEND=true
+  fi
+done
+
+# This updates the lockfile without building
+cargo metadata --quiet > /dev/null
+
+pushd nodejs || exit 1
+npm install --package-lock-only --silent
+popd
+pushd node || exit 1
+npm install --package-lock-only --silent
+popd
+
+if git diff --quiet --exit-code; then
+  echo "No lockfile changes to commit; skipping amend."
+elif $AMEND; then
+  git add Cargo.lock nodejs/package-lock.json node/package-lock.json
+  git commit --amend --no-edit
+else
+  git add Cargo.lock nodejs/package-lock.json node/package-lock.json
+  git commit -m "Update lockfiles"
+fi

docs/README.md

@@ -2,7 +2,7 @@
 
 LanceDB docs are deployed to https://lancedb.github.io/lancedb/.
 
-Docs is built and deployed automatically by [Github Actions](.github/workflows/docs.yml)
+Docs is built and deployed automatically by [Github Actions](../.github/workflows/docs.yml)
 whenever a commit is pushed to the `main` branch. So it is possible for the docs to show
 unreleased features.
 

docs/mkdocs.yml

@@ -193,6 +193,7 @@ nav:
           - Pandas and PyArrow: python/pandas_and_pyarrow.md
           - Polars: python/polars_arrow.md
           - DuckDB: python/duckdb.md
+          - Datafusion: python/datafusion.md
           - LangChain:
               - LangChain 🔗: integrations/langchain.md
               - LangChain demo: notebooks/langchain_demo.ipynb
@@ -205,6 +206,7 @@ nav:
           - PromptTools: integrations/prompttools.md
           - dlt: integrations/dlt.md
           - phidata: integrations/phidata.md
+          - Genkit: integrations/genkit.md
       - 🎯 Examples:
           - Overview: examples/index.md
           - 🐍 Python:
@@ -247,6 +249,7 @@ nav:
       - Data management: concepts/data_management.md
   - Guides:
       - Working with tables: guides/tables.md
+      - Working with SQL: guides/sql_querying.md
       - Building an ANN index: ann_indexes.md
       - Vector Search: search.md
       - Full-text search (native): fts.md
@@ -323,6 +326,7 @@ nav:
       - Pandas and PyArrow: python/pandas_and_pyarrow.md
       - Polars: python/polars_arrow.md
       - DuckDB: python/duckdb.md
+      - Datafusion: python/datafusion.md
       - LangChain 🦜️🔗↗: integrations/langchain.md
       - LangChain.js 🦜️🔗↗: https://js.langchain.com/docs/integrations/vectorstores/lancedb
       - LlamaIndex 🦙↗: integrations/llamaIndex.md
@@ -331,6 +335,7 @@ nav:
       - PromptTools: integrations/prompttools.md
       - dlt: integrations/dlt.md
       - phidata: integrations/phidata.md
+      - Genkit: integrations/genkit.md
   - Examples:
       - examples/index.md
       - 🐍 Python:

docs/overrides/partials/main.html

@@ -0,0 +1,5 @@
+{% extends "base.html" %}
+
+{% block announce %}
+  📚 Starting June 1st, 2025, please use <a href="https://lancedb.github.io/documentation" target="_blank" rel="noopener noreferrer">lancedb.github.io/documentation</a> for the latest docs.
+{% endblock %}

docs/src/ann_indexes.md

@@ -291,7 +291,7 @@ Product quantization can lead to approximately `16 * sizeof(float32) / 1 = 64` t
 
 `num_partitions` is used to decide how many partitions the first level `IVF` index uses.
 Higher number of partitions could lead to more efficient I/O during queries and better accuracy, but it takes much more time to train.
-On `SIFT-1M` dataset, our benchmark shows that keeping each partition 1K-4K rows lead to a good latency / recall.
+On `SIFT-1M` dataset, our benchmark shows that keeping each partition 4K-8K rows lead to a good latency / recall.
 
 `num_sub_vectors` specifies how many Product Quantization (PQ) short codes to generate on each vector. The number should be a factor of the vector dimension. Because
 PQ is a lossy compression of the original vector, a higher `num_sub_vectors` usually results in

docs/src/guides/sql_querying.md

@@ -0,0 +1,68 @@
+You can use DuckDB and Apache Datafusion to query your LanceDB tables using SQL.
+This guide will show how to query Lance tables them using both.
+
+We will re-use the dataset [created previously](./pandas_and_pyarrow.md):
+
+```python
+import lancedb
+
+db = lancedb.connect("data/sample-lancedb")
+data = [
+    {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+    {"vector": [5.9, 26.5], "item": "bar", "price": 20.0}
+]
+table = db.create_table("pd_table", data=data)
+```
+
+## Querying a LanceDB Table with DuckDb
+
+The `to_lance` method converts the LanceDB table to a `LanceDataset`, which is accessible to DuckDB through the Arrow compatibility layer.
+To query the resulting Lance dataset in DuckDB, all you need to do is reference the dataset by the same name in your SQL query.
+
+```python
+import duckdb
+
+arrow_table = table.to_lance()
+
+duckdb.query("SELECT * FROM arrow_table")
+```
+
+```
+┌─────────────┬─────────┬────────┐
+│   vector    │  item   │ price  │
+│   float[]   │ varchar │ double │
+├─────────────┼─────────┼────────┤
+│ [3.1, 4.1]  │ foo     │   10.0 │
+│ [5.9, 26.5] │ bar     │   20.0 │
+└─────────────┴─────────┴────────┘
+```
+
+## Querying a LanceDB Table with Apache Datafusion
+
+Have the required imports before doing any querying.
+
+=== "Python"
+
+    ```python
+    --8<-- "python/python/tests/docs/test_guide_tables.py:import-lancedb"
+    --8<-- "python/python/tests/docs/test_guide_tables.py:import-session-context"
+    --8<-- "python/python/tests/docs/test_guide_tables.py:import-ffi-dataset"
+    ```
+
+Register the table created with the Datafusion session context.
+
+=== "Python"
+
+    ```python
+    --8<-- "python/python/tests/docs/test_guide_tables.py:lance_sql_basic"
+    ```
+
+```
+┌─────────────┬─────────┬────────┐
+│   vector    │  item   │ price  │
+│   float[]   │ varchar │ double │
+├─────────────┼─────────┼────────┤
+│ [3.1, 4.1]  │ foo     │   10.0 │
+│ [5.9, 26.5] │ bar     │   20.0 │
+└─────────────┴─────────┴────────┘
+```

docs/src/guides/storage.md

@@ -342,7 +342,7 @@ For **read and write access**, LanceDB will need a policy such as:
             "Action": [
               "s3:PutObject",
               "s3:GetObject",
-              "s3:DeleteObject",
+              "s3:DeleteObject"
             ],
             "Resource": "arn:aws:s3:::<bucket>/<prefix>/*"
         },
@@ -374,7 +374,7 @@ For **read-only access**, LanceDB will need a policy such as:
         {
             "Effect": "Allow",
             "Action": [
-              "s3:GetObject",
+              "s3:GetObject"
             ],
             "Resource": "arn:aws:s3:::<bucket>/<prefix>/*"
         },

docs/src/guides/tables.md

@@ -765,7 +765,10 @@ This can be used to update zero to all rows depending on how many rows match the
         ];
         const tbl = await db.createTable("my_table", data)
 
-        await tbl.update({vector: [10, 10]}, { where: "x = 2"})
+        await tbl.update({
+            values: { vector: [10, 10] },
+            where: "x = 2"
+        });
         ```
 
     === "vectordb (deprecated)"
@@ -784,7 +787,10 @@ This can be used to update zero to all rows depending on how many rows match the
         ];
         const tbl = await db.createTable("my_table", data)
 
-        await tbl.update({ where: "x = 2", values: {vector: [10, 10]} })
+        await tbl.update({
+            where: "x = 2",
+            values: { vector: [10, 10] }
+        });
         ```
 
 #### Updating using a sql query

docs/src/integrations/genkit.md

@@ -0,0 +1,183 @@
+### genkitx-lancedb
+This is a lancedb plugin for genkit framework. It allows you to use LanceDB for ingesting and rereiving data using genkit framework.
+
+![integration-banner-genkit](https://github.com/user-attachments/assets/a6cc28af-98e9-4425-b87c-7ab139bd7893)
+
+### Installation
+```bash
+pnpm install genkitx-lancedb
+```
+
+### Usage
+
+Adding LanceDB plugin to your genkit instance.
+
+```ts
+import { lancedbIndexerRef, lancedb, lancedbRetrieverRef, WriteMode } from 'genkitx-lancedb';
+import { textEmbedding004, vertexAI } from '@genkit-ai/vertexai';
+import { gemini } from '@genkit-ai/vertexai';
+import { z, genkit } from 'genkit';
+import { Document } from 'genkit/retriever';
+import { chunk } from 'llm-chunk';
+import { readFile } from 'fs/promises';
+import path from 'path';
+import pdf from 'pdf-parse/lib/pdf-parse';
+
+const ai = genkit({
+  plugins: [
+    // vertexAI provides the textEmbedding004 embedder
+    vertexAI(),
+
+    // the local vector store requires an embedder to translate from text to vector
+    lancedb([
+      {
+        dbUri: '.db', // optional lancedb uri, default to .db
+        tableName: 'table', // optional table name, default to table
+        embedder: textEmbedding004,
+      },
+    ]),
+  ],
+});
+```
+
+You can run this app with the following command:
+```bash
+genkit start -- tsx --watch src/index.ts
+```
+
+This'll add LanceDB as a retriever and indexer to the genkit instance. You can see it in the GUI view
+<img width="1710" alt="Screenshot 2025-05-11 at 7 21 05 PM" src="https://github.com/user-attachments/assets/e752f7f4-785b-4797-a11e-72ab06a531b7" />
+
+**Testing retrieval on a sample table**
+Let's see the raw retrieval results
+
+<img width="1710" alt="Screenshot 2025-05-11 at 7 21 05 PM" src="https://github.com/user-attachments/assets/b8d356ed-8421-4790-8fc0-d6af563b9657" />
+On running this query, you'll 5 results fetched from the lancedb table, where each result looks something like this:
+<img width="1417" alt="Screenshot 2025-05-11 at 7 21 18 PM" src="https://github.com/user-attachments/assets/77429525-36e2-4da6-a694-e58c1cf9eb83" />
+
+
+
+## Creating a custom RAG flow
+
+Now that we've seen how you can use LanceDB for in a genkit pipeline, let's refine the flow and create a RAG. A RAG flow will consist of an index and a retreiver with its outputs postprocessed an fed into an LLM for final response
+
+### Creating custom indexer flows
+You can also create custom indexer flows, utilizing more options and features provided by LanceDB.
+
+```ts
+export const menuPdfIndexer = lancedbIndexerRef({
+   // Using all defaults, for dbUri, tableName, and embedder, etc
+});
+
+const chunkingConfig = {
+  minLength: 1000,
+  maxLength: 2000,
+  splitter: 'sentence',
+  overlap: 100,
+  delimiters: '',
+} as any;
+
+
+async function extractTextFromPdf(filePath: string) {
+  const pdfFile = path.resolve(filePath);
+  const dataBuffer = await readFile(pdfFile);
+  const data = await pdf(dataBuffer);
+  return data.text;
+}
+
+export const indexMenu = ai.defineFlow(
+  {
+    name: 'indexMenu',
+    inputSchema: z.string().describe('PDF file path'),
+    outputSchema: z.void(),
+  },
+  async (filePath: string) => {
+    filePath = path.resolve(filePath);
+
+    // Read the pdf.
+    const pdfTxt = await ai.run('extract-text', () =>
+      extractTextFromPdf(filePath)
+    );
+
+    // Divide the pdf text into segments.
+    const chunks = await ai.run('chunk-it', async () =>
+      chunk(pdfTxt, chunkingConfig)
+    );
+
+    // Convert chunks of text into documents to store in the index.
+    const documents = chunks.map((text) => {
+      return Document.fromText(text, { filePath });
+    });
+
+    // Add documents to the index.
+    await ai.index({
+      indexer: menuPdfIndexer,
+      documents,
+      options: {
+        writeMode: WriteMode.Overwrite,
+      } as any
+    });
+  }
+);
+```
+
+<img width="1316" alt="Screenshot 2025-05-11 at 8 35 56 PM" src="https://github.com/user-attachments/assets/e2a20ce4-d1d0-4fa2-9a84-f2cc26e3a29f" />
+
+In your console, you can see the logs
+
+<img width="511" alt="Screenshot 2025-05-11 at 7 19 14 PM" src="https://github.com/user-attachments/assets/243f26c5-ed38-40b6-b661-002f40f0423a" />
+
+### Creating custom retriever flows
+You can also create custom retriever flows, utilizing more options and features provided by LanceDB.
+```ts
+export const menuRetriever = lancedbRetrieverRef({
+  tableName: "table", // Use the same table name as the indexer.
+  displayName: "Menu", // Use a custom display name.
+
+export const menuQAFlow = ai.defineFlow(
+  { name: "Menu", inputSchema: z.string(), outputSchema: z.string() },
+  async (input: string) => {
+    // retrieve relevant documents
+    const docs = await ai.retrieve({
+      retriever: menuRetriever,
+      query: input,
+      options: { 
+        k: 3,
+      },
+    });
+
+    const extractedContent = docs.map(doc => {
+      if (doc.content && Array.isArray(doc.content) && doc.content.length > 0) {
+        if (doc.content[0].media && doc.content[0].media.url) {
+          return doc.content[0].media.url;
+        }
+      }
+      return "No content found";
+    });
+
+    console.log("Extracted content:", extractedContent);
+
+    const { text } = await ai.generate({
+      model: gemini('gemini-2.0-flash'),
+      prompt: `
+You are acting as a helpful AI assistant that can answer 
+questions about the food available on the menu at Genkit Grub Pub.
+
+Use only the context provided to answer the question.
+If you don't know, do not make up an answer.
+Do not add or change items on the menu.
+
+Context:
+${extractedContent.join('\n\n')}
+
+Question: ${input}`,
+      docs,
+    });
+    
+    return text;
+  }
+);
+```
+Now using our retrieval flow, we can ask question about the ingsted PDF
+<img width="1306" alt="Screenshot 2025-05-11 at 7 18 45 PM" src="https://github.com/user-attachments/assets/86c66b13-7c12-4d5f-9d81-ae36bfb1c346" />
+

docs/src/js/classes/BooleanQuery.md

@@ -0,0 +1,53 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / BooleanQuery
+
+# Class: BooleanQuery
+
+Represents a full-text query interface.
+This interface defines the structure and behavior for full-text queries,
+including methods to retrieve the query type and convert the query to a dictionary format.
+
+## Implements
+
+- [`FullTextQuery`](../interfaces/FullTextQuery.md)
+
+## Constructors
+
+### new BooleanQuery()
+
+```ts
+new BooleanQuery(queries): BooleanQuery
+```
+
+Creates an instance of BooleanQuery.
+
+#### Parameters
+
+* **queries**: [[`Occur`](../enumerations/Occur.md), [`FullTextQuery`](../interfaces/FullTextQuery.md)][]
+    An array of (Occur, FullTextQuery objects) to combine.
+    Occur specifies whether the query must match, or should match.
+
+#### Returns
+
+[`BooleanQuery`](BooleanQuery.md)
+
+## Methods
+
+### queryType()
+
+```ts
+queryType(): FullTextQueryType
+```
+
+The type of the full-text query.
+
+#### Returns
+
+[`FullTextQueryType`](../enumerations/FullTextQueryType.md)
+
+#### Implementation of
+
+[`FullTextQuery`](../interfaces/FullTextQuery.md).[`queryType`](../interfaces/FullTextQuery.md#querytype)

docs/src/js/classes/MatchQuery.md

@@ -40,6 +40,7 @@ Creates an instance of MatchQuery.
     - `boost`: The boost factor for the query (default is 1.0).
     - `fuzziness`: The fuzziness level for the query (default is 0).
     - `maxExpansions`: The maximum number of terms to consider for fuzzy matching (default is 50).
+    - `operator`: The logical operator to use for combining terms in the query (default is "OR").
 
 * **options.boost?**: `number`
 
@@ -47,6 +48,8 @@ Creates an instance of MatchQuery.
 
 * **options.maxExpansions?**: `number`
 
+* **options.operator?**: [`Operator`](../enumerations/Operator.md)
+
 #### Returns
 
 [`MatchQuery`](MatchQuery.md)

docs/src/js/classes/MergeInsertBuilder.md

@@ -33,20 +33,22 @@ Construct a MergeInsertBuilder. __Internal use only.__
 ### execute()
 
 ```ts
-execute(data): Promise<void>
+execute(data, execOptions?): Promise<MergeResult>
 ```
 
 Executes the merge insert operation
 
-Nothing is returned but the `Table` is updated
-
 #### Parameters
 
 * **data**: [`Data`](../type-aliases/Data.md)
 
+* **execOptions?**: `Partial`&lt;[`WriteExecutionOptions`](../interfaces/WriteExecutionOptions.md)&gt;
+
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`MergeResult`](../interfaces/MergeResult.md)&gt;
+
+the merge result
 
 ***
 

docs/src/js/classes/MultiMatchQuery.md

@@ -38,9 +38,12 @@ Creates an instance of MultiMatchQuery.
 * **options?**
     Optional parameters for the multi-match query.
     - `boosts`: An array of boost factors for each column (default is 1.0 for all).
+    - `operator`: The logical operator to use for combining terms in the query (default is "OR").
 
 * **options.boosts?**: `number`[]
 
+* **options.operator?**: [`Operator`](../enumerations/Operator.md)
+
 #### Returns
 
 [`MultiMatchQuery`](MultiMatchQuery.md)

docs/src/js/classes/PhraseQuery.md

@@ -19,7 +19,10 @@ including methods to retrieve the query type and convert the query to a dictiona
 ### new PhraseQuery()
 
 ```ts
-new PhraseQuery(query, column): PhraseQuery
+new PhraseQuery(
+   query,
+   column,
+   options?): PhraseQuery
 ```
 
 Creates an instance of `PhraseQuery`.
@@ -32,6 +35,12 @@ Creates an instance of `PhraseQuery`.
 * **column**: `string`
     The name of the column to search within.
 
+* **options?**
+    Optional parameters for the phrase query.
+    - `slop`: The maximum number of intervening unmatched positions allowed between words in the phrase (default is 0).
+
+* **options.slop?**: `number`
+
 #### Returns
 
 [`PhraseQuery`](PhraseQuery.md)

docs/src/js/classes/Table.md

@@ -40,7 +40,7 @@ Returns the name of the table
 ### add()
 
 ```ts
-abstract add(data, options?): Promise<void>
+abstract add(data, options?): Promise<AddResult>
 ```
 
 Insert records into this Table.
@@ -54,14 +54,17 @@ Insert records into this Table.
 
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`AddResult`](../interfaces/AddResult.md)&gt;
+
+A promise that resolves to an object
+containing the new version number of the table
 
 ***
 
 ### addColumns()
 
 ```ts
-abstract addColumns(newColumnTransforms): Promise<void>
+abstract addColumns(newColumnTransforms): Promise<AddColumnsResult>
 ```
 
 Add new columns with defined values.
@@ -76,14 +79,17 @@ Add new columns with defined values.
 
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`AddColumnsResult`](../interfaces/AddColumnsResult.md)&gt;
+
+A promise that resolves to an object
+containing the new version number of the table after adding the columns.
 
 ***
 
 ### alterColumns()
 
 ```ts
-abstract alterColumns(columnAlterations): Promise<void>
+abstract alterColumns(columnAlterations): Promise<AlterColumnsResult>
 ```
 
 Alter the name or nullability of columns.
@@ -96,7 +102,10 @@ Alter the name or nullability of columns.
 
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`AlterColumnsResult`](../interfaces/AlterColumnsResult.md)&gt;
+
+A promise that resolves to an object
+containing the new version number of the table after altering the columns.
 
 ***
 
@@ -117,8 +126,8 @@ wish to return to standard mode, call `checkoutLatest`.
 
 #### Parameters
 
-* **version**: `number`
+* **version**: `string` \| `number`
-    The version to checkout
+    The version to checkout, could be version number or tag
 
 #### Returns
 
@@ -252,7 +261,7 @@ await table.createIndex("my_float_col");
 ### delete()
 
 ```ts
-abstract delete(predicate): Promise<void>
+abstract delete(predicate): Promise<DeleteResult>
 ```
 
 Delete the rows that satisfy the predicate.
@@ -263,7 +272,10 @@ Delete the rows that satisfy the predicate.
 
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`DeleteResult`](../interfaces/DeleteResult.md)&gt;
+
+A promise that resolves to an object
+containing the new version number of the table
 
 ***
 
@@ -284,7 +296,7 @@ Return a brief description of the table
 ### dropColumns()
 
 ```ts
-abstract dropColumns(columnNames): Promise<void>
+abstract dropColumns(columnNames): Promise<DropColumnsResult>
 ```
 
 Drop one or more columns from the dataset
@@ -303,7 +315,10 @@ then call ``cleanup_files`` to remove the old files.
 
 #### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`DropColumnsResult`](../interfaces/DropColumnsResult.md)&gt;
+
+A promise that resolves to an object
+containing the new version number of the table after dropping the columns.
 
 ***
 
@@ -454,6 +469,28 @@ Modeled after ``VACUUM`` in PostgreSQL.
 
 ***
 
+### prewarmIndex()
+
+```ts
+abstract prewarmIndex(name): Promise<void>
+```
+
+Prewarm an index in the table.
+
+#### Parameters
+
+* **name**: `string`
+    The name of the index.
+    This will load the index into memory.  This may reduce the cold-start time for
+    future queries.  If the index does not fit in the cache then this call may be
+    wasteful.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
 ### query()
 
 ```ts
@@ -593,6 +630,50 @@ of the given query
 
 ***
 
+### stats()
+
+```ts
+abstract stats(): Promise<TableStatistics>
+```
+
+Returns table and fragment statistics
+
+#### Returns
+
+`Promise`&lt;[`TableStatistics`](../interfaces/TableStatistics.md)&gt;
+
+The table and fragment statistics
+
+***
+
+### tags()
+
+```ts
+abstract tags(): Promise<Tags>
+```
+
+Get a tags manager for this table.
+
+Tags allow you to label specific versions of a table with a human-readable name.
+The returned tags manager can be used to list, create, update, or delete tags.
+
+#### Returns
+
+`Promise`&lt;[`Tags`](Tags.md)&gt;
+
+A tags manager for this table
+
+#### Example
+
+```typescript
+const tagsManager = await table.tags();
+await tagsManager.create("v1", 1);
+const tags = await tagsManager.list();
+console.log(tags); // { "v1": { version: 1, manifestSize: ... } }
+```
+
+***
+
 ### toArrow()
 
 ```ts
@@ -612,7 +693,7 @@ Return the table as an arrow table
 #### update(opts)
 
 ```ts
-abstract update(opts): Promise<void>
+abstract update(opts): Promise<UpdateResult>
 ```
 
 Update existing records in the Table
@@ -623,7 +704,10 @@ Update existing records in the Table
 
 ##### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`UpdateResult`](../interfaces/UpdateResult.md)&gt;
+
+A promise that resolves to an object containing
+the number of rows updated and the new version number
 
 ##### Example
 
@@ -634,7 +718,7 @@ table.update({where:"x = 2", values:{"vector": [10, 10]}})
 #### update(opts)
 
 ```ts
-abstract update(opts): Promise<void>
+abstract update(opts): Promise<UpdateResult>
 ```
 
 Update existing records in the Table
@@ -645,7 +729,10 @@ Update existing records in the Table
 
 ##### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`UpdateResult`](../interfaces/UpdateResult.md)&gt;
+
+A promise that resolves to an object containing
+the number of rows updated and the new version number
 
 ##### Example
 
@@ -656,7 +743,7 @@ table.update({where:"x = 2", valuesSql:{"x": "x + 1"}})
 #### update(updates, options)
 
 ```ts
-abstract update(updates, options?): Promise<void>
+abstract update(updates, options?): Promise<UpdateResult>
 ```
 
 Update existing records in the Table
@@ -679,10 +766,6 @@ repeatedly calilng this method.
 * **updates**: `Record`&lt;`string`, `string`&gt; \| `Map`&lt;`string`, `string`&gt;
     the
     columns to update
-    Keys in the map should specify the name of the column to update.
-    Values in the map provide the new value of the column.  These can
-    be SQL literal strings (e.g. "7" or "'foo'") or they can be expressions
-    based on the row being updated (e.g. "my_col + 1")
 
 * **options?**: `Partial`&lt;[`UpdateOptions`](../interfaces/UpdateOptions.md)&gt;
     additional options to control
@@ -690,7 +773,15 @@ repeatedly calilng this method.
 
 ##### Returns
 
-`Promise`&lt;`void`&gt;
+`Promise`&lt;[`UpdateResult`](../interfaces/UpdateResult.md)&gt;
+
+A promise that resolves to an object
+containing the number of rows updated and the new version number
+
+Keys in the map should specify the name of the column to update.
+Values in the map provide the new value of the column.  These can
+be SQL literal strings (e.g. "7" or "'foo'") or they can be expressions
+based on the row being updated (e.g. "my_col + 1")
 
 ***
 
@@ -731,3 +822,26 @@ Retrieve the version of the table
 #### Returns
 
 `Promise`&lt;`number`&gt;
+
+***
+
+### waitForIndex()
+
+```ts
+abstract waitForIndex(indexNames, timeoutSeconds): Promise<void>
+```
+
+Waits for asynchronous indexing to complete on the table.
+
+#### Parameters
+
+* **indexNames**: `string`[]
+    The name of the indices to wait for
+
+* **timeoutSeconds**: `number`
+    The number of seconds to wait before timing out
+    This will raise an error if the indices are not created and fully indexed within the timeout.
+
+#### Returns
+
+`Promise`&lt;`void`&gt;

docs/src/js/classes/TagContents.md

@@ -0,0 +1,35 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / TagContents
+
+# Class: TagContents
+
+## Constructors
+
+### new TagContents()
+
+```ts
+new TagContents(): TagContents
+```
+
+#### Returns
+
+[`TagContents`](TagContents.md)
+
+## Properties
+
+### manifestSize
+
+```ts
+manifestSize: number;
+```
+
+***
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/classes/Tags.md

@@ -0,0 +1,99 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / Tags
+
+# Class: Tags
+
+## Constructors
+
+### new Tags()
+
+```ts
+new Tags(): Tags
+```
+
+#### Returns
+
+[`Tags`](Tags.md)
+
+## Methods
+
+### create()
+
+```ts
+create(tag, version): Promise<void>
+```
+
+#### Parameters
+
+* **tag**: `string`
+
+* **version**: `number`
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
+### delete()
+
+```ts
+delete(tag): Promise<void>
+```
+
+#### Parameters
+
+* **tag**: `string`
+
+#### Returns
+
+`Promise`&lt;`void`&gt;
+
+***
+
+### getVersion()
+
+```ts
+getVersion(tag): Promise<number>
+```
+
+#### Parameters
+
+* **tag**: `string`
+
+#### Returns
+
+`Promise`&lt;`number`&gt;
+
+***
+
+### list()
+
+```ts
+list(): Promise<Record<string, TagContents>>
+```
+
+#### Returns
+
+`Promise`&lt;`Record`&lt;`string`, [`TagContents`](TagContents.md)&gt;&gt;
+
+***
+
+### update()
+
+```ts
+update(tag, version): Promise<void>
+```
+
+#### Parameters
+
+* **tag**: `string`
+
+* **version**: `number`
+
+#### Returns
+
+`Promise`&lt;`void`&gt;

docs/src/js/enumerations/FullTextQueryType.md

@@ -15,6 +15,14 @@ Enum representing the types of full-text queries supported.
 
 ## Enumeration Members
 
+### Boolean
+
+```ts
+Boolean: "boolean";
+```
+
+***
+
 ### Boost
 
 ```ts

docs/src/js/enumerations/Occur.md

@@ -0,0 +1,28 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / Occur
+
+# Enumeration: Occur
+
+Enum representing the occurrence of terms in full-text queries.
+
+- `Must`: The term must be present in the document.
+- `Should`: The term should contribute to the document score, but is not required.
+
+## Enumeration Members
+
+### Must
+
+```ts
+Must: "MUST";
+```
+
+***
+
+### Should
+
+```ts
+Should: "SHOULD";
+```

docs/src/js/enumerations/Operator.md

@@ -0,0 +1,28 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / Operator
+
+# Enumeration: Operator
+
+Enum representing the logical operators used in full-text queries.
+
+- `And`: All terms must match.
+- `Or`: At least one term must match.
+
+## Enumeration Members
+
+### And
+
+```ts
+And: "AND";
+```
+
+***
+
+### Or
+
+```ts
+Or: "OR";
+```

docs/src/js/globals.md

@@ -12,9 +12,12 @@
 ## Enumerations
 
 - [FullTextQueryType](enumerations/FullTextQueryType.md)
+- [Occur](enumerations/Occur.md)
+- [Operator](enumerations/Operator.md)
 
 ## Classes
 
+- [BooleanQuery](classes/BooleanQuery.md)
 - [BoostQuery](classes/BoostQuery.md)
 - [Connection](classes/Connection.md)
 - [Index](classes/Index.md)
@@ -27,19 +30,28 @@
 - [QueryBase](classes/QueryBase.md)
 - [RecordBatchIterator](classes/RecordBatchIterator.md)
 - [Table](classes/Table.md)
+- [TagContents](classes/TagContents.md)
+- [Tags](classes/Tags.md)
 - [VectorColumnOptions](classes/VectorColumnOptions.md)
 - [VectorQuery](classes/VectorQuery.md)
 
 ## Interfaces
 
+- [AddColumnsResult](interfaces/AddColumnsResult.md)
 - [AddColumnsSql](interfaces/AddColumnsSql.md)
 - [AddDataOptions](interfaces/AddDataOptions.md)
+- [AddResult](interfaces/AddResult.md)
+- [AlterColumnsResult](interfaces/AlterColumnsResult.md)
 - [ClientConfig](interfaces/ClientConfig.md)
 - [ColumnAlteration](interfaces/ColumnAlteration.md)
 - [CompactionStats](interfaces/CompactionStats.md)
 - [ConnectionOptions](interfaces/ConnectionOptions.md)
 - [CreateTableOptions](interfaces/CreateTableOptions.md)
+- [DeleteResult](interfaces/DeleteResult.md)
+- [DropColumnsResult](interfaces/DropColumnsResult.md)
 - [ExecutableQuery](interfaces/ExecutableQuery.md)
+- [FragmentStatistics](interfaces/FragmentStatistics.md)
+- [FragmentSummaryStats](interfaces/FragmentSummaryStats.md)
 - [FtsOptions](interfaces/FtsOptions.md)
 - [FullTextQuery](interfaces/FullTextQuery.md)
 - [FullTextSearchOptions](interfaces/FullTextSearchOptions.md)
@@ -50,6 +62,7 @@
 - [IndexStatistics](interfaces/IndexStatistics.md)
 - [IvfFlatOptions](interfaces/IvfFlatOptions.md)
 - [IvfPqOptions](interfaces/IvfPqOptions.md)
+- [MergeResult](interfaces/MergeResult.md)
 - [OpenTableOptions](interfaces/OpenTableOptions.md)
 - [OptimizeOptions](interfaces/OptimizeOptions.md)
 - [OptimizeStats](interfaces/OptimizeStats.md)
@@ -57,9 +70,12 @@
 - [RemovalStats](interfaces/RemovalStats.md)
 - [RetryConfig](interfaces/RetryConfig.md)
 - [TableNamesOptions](interfaces/TableNamesOptions.md)
+- [TableStatistics](interfaces/TableStatistics.md)
 - [TimeoutConfig](interfaces/TimeoutConfig.md)
 - [UpdateOptions](interfaces/UpdateOptions.md)
+- [UpdateResult](interfaces/UpdateResult.md)
 - [Version](interfaces/Version.md)
+- [WriteExecutionOptions](interfaces/WriteExecutionOptions.md)
 
 ## Type Aliases
 

docs/src/js/interfaces/AddColumnsResult.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / AddColumnsResult
+
+# Interface: AddColumnsResult
+
+## Properties
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/AddResult.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / AddResult
+
+# Interface: AddResult
+
+## Properties
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/AlterColumnsResult.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / AlterColumnsResult
+
+# Interface: AlterColumnsResult
+
+## Properties
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/DeleteResult.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DeleteResult
+
+# Interface: DeleteResult
+
+## Properties
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/DropColumnsResult.md

@@ -0,0 +1,15 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / DropColumnsResult
+
+# Interface: DropColumnsResult
+
+## Properties
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/FragmentStatistics.md

@@ -0,0 +1,37 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / FragmentStatistics
+
+# Interface: FragmentStatistics
+
+## Properties
+
+### lengths
+
+```ts
+lengths: FragmentSummaryStats;
+```
+
+Statistics on the number of rows in the table fragments
+
+***
+
+### numFragments
+
+```ts
+numFragments: number;
+```
+
+The number of fragments in the table
+
+***
+
+### numSmallFragments
+
+```ts
+numSmallFragments: number;
+```
+
+The number of uncompacted fragments in the table

docs/src/js/interfaces/FragmentSummaryStats.md

@@ -0,0 +1,77 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / FragmentSummaryStats
+
+# Interface: FragmentSummaryStats
+
+## Properties
+
+### max
+
+```ts
+max: number;
+```
+
+The number of rows in the fragment with the most rows
+
+***
+
+### mean
+
+```ts
+mean: number;
+```
+
+The mean number of rows in the fragments
+
+***
+
+### min
+
+```ts
+min: number;
+```
+
+The number of rows in the fragment with the fewest rows
+
+***
+
+### p25
+
+```ts
+p25: number;
+```
+
+The 25th percentile of number of rows in the fragments
+
+***
+
+### p50
+
+```ts
+p50: number;
+```
+
+The 50th percentile of number of rows in the fragments
+
+***
+
+### p75
+
+```ts
+p75: number;
+```
+
+The 75th percentile of number of rows in the fragments
+
+***
+
+### p99
+
+```ts
+p99: number;
+```
+
+The 99th percentile of number of rows in the fragments

docs/src/js/interfaces/IndexOptions.md

@@ -39,3 +39,11 @@ and the same name, then an error will be returned.  This is true even if
 that index is out of date.
 
 The default is true
+
+***
+
+### waitTimeoutSeconds?
+
+```ts
+optional waitTimeoutSeconds: number;
+```

docs/src/js/interfaces/MergeResult.md

@@ -0,0 +1,39 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / MergeResult
+
+# Interface: MergeResult
+
+## Properties
+
+### numDeletedRows
+
+```ts
+numDeletedRows: number;
+```
+
+***
+
+### numInsertedRows
+
+```ts
+numInsertedRows: number;
+```
+
+***
+
+### numUpdatedRows
+
+```ts
+numUpdatedRows: number;
+```
+
+***
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/TableStatistics.md

@@ -0,0 +1,47 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / TableStatistics
+
+# Interface: TableStatistics
+
+## Properties
+
+### fragmentStats
+
+```ts
+fragmentStats: FragmentStatistics;
+```
+
+Statistics on table fragments
+
+***
+
+### numIndices
+
+```ts
+numIndices: number;
+```
+
+The number of indices in the table
+
+***
+
+### numRows
+
+```ts
+numRows: number;
+```
+
+The number of rows in the table
+
+***
+
+### totalBytes
+
+```ts
+totalBytes: number;
+```
+
+The total number of bytes in the table

docs/src/js/interfaces/UpdateResult.md

@@ -0,0 +1,23 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / UpdateResult
+
+# Interface: UpdateResult
+
+## Properties
+
+### rowsUpdated
+
+```ts
+rowsUpdated: number;
+```
+
+***
+
+### version
+
+```ts
+version: number;
+```

docs/src/js/interfaces/WriteExecutionOptions.md

@@ -0,0 +1,26 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / WriteExecutionOptions
+
+# Interface: WriteExecutionOptions
+
+## Properties
+
+### timeoutMs?
+
+```ts
+optional timeoutMs: number;
+```
+
+Maximum time to run the operation before cancelling it.
+
+By default, there is a 30-second timeout that is only enforced after the
+first attempt. This is to prevent spending too long retrying to resolve
+conflicts. For example, if a write attempt takes 20 seconds and fails,
+the second attempt will be cancelled after 10 seconds, hitting the
+30-second timeout. However, a write that takes one hour and succeeds on the
+first attempt will not be cancelled.
+
+When this is set, the timeout is enforced on all attempts, including the first.

docs/src/python/datafusion.md

@@ -0,0 +1,53 @@
+# Apache Datafusion
+
+In Python, LanceDB tables can also be queried with [Apache Datafusion](https://datafusion.apache.org/), an extensible query engine written in Rust that uses Apache Arrow as its in-memory format. This means you can write complex SQL queries to analyze your data in LanceDB.
+
+This integration is done via [Datafusion FFI](https://docs.rs/datafusion-ffi/latest/datafusion_ffi/), which provides a native integration between LanceDB and Datafusion.
+The Datafusion FFI allows to pass down column selections and basic filters to LanceDB, reducing the amount of scanned data when executing your query. Additionally, the integration allows streaming data from LanceDB tables which allows to do aggregation larger-than-memory.
+
+We can demonstrate this by first installing `datafusion` and `lancedb`.
+
+```shell
+pip install datafusion lancedb
+```
+
+We will re-use the dataset [created previously](./pandas_and_pyarrow.md):
+
+```python
+import lancedb
+
+from datafusion import SessionContext
+from lance import FFILanceTableProvider
+
+db = lancedb.connect("data/sample-lancedb")
+data = [
+    {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+    {"vector": [5.9, 26.5], "item": "bar", "price": 20.0}
+]
+lance_table = db.create_table("lance_table", data)
+
+ctx = SessionContext()
+
+ffi_lance_table = FFILanceTableProvider(
+    lance_table.to_lance(), with_row_id=True, with_row_addr=True
+)
+ctx.register_table_provider("ffi_lance_table", ffi_lance_table)
+```
+
+The `to_lance` method converts the LanceDB table to a `LanceDataset`, which is accessible to Datafusion through the Datafusion FFI integration layer.
+To query the resulting Lance dataset in Datafusion, you first need to register the dataset with Datafusion and then just reference it by the same name in your SQL query.
+
+```python
+ctx.table("ffi_lance_table")
+ctx.sql("SELECT * FROM ffi_lance_table")
+```
+
+```
+┌─────────────┬─────────┬────────┬─────────────────┬─────────────────┐
+│   vector    │  item   │ price  │ _rowid          │ _rowaddr        │
+│   float[]   │ varchar │ double │ bigint unsigned │ bigint unsigned │
+├─────────────┼─────────┼────────┼─────────────────┼─────────────────┤
+│ [3.1, 4.1]  │ foo     │   10.0 │               0 │               0 │
+│ [5.9, 26.5] │ bar     │   20.0 │               1 │               1 │
+└─────────────┴─────────┴────────┴─────────────────┴─────────────────┘
+```

docs/test/requirements.txt

@@ -7,3 +7,4 @@ tantivy==0.20.1
 --extra-index-url https://download.pytorch.org/whl/cpu
 torch
 polars>=0.19, <=1.3.0
+datafusion

java/core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
         <groupId>com.lancedb</groupId>
         <artifactId>lancedb-parent</artifactId>
-        <version>0.19.0-beta.6</version>
+        <version>0.20.1-beta.2</version>
         <relativePath>../pom.xml</relativePath>
     </parent>
 

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.19.0-beta.6</version>
+    <version>0.20.1-beta.2</version>
     <packaging>pom</packaging>
 
     <name>LanceDB Parent</name>

node/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "vectordb",
-  "version": "0.19.0-beta.6",
+  "version": "0.20.1-beta.2",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "vectordb",
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
       "cpu": [
         "x64",
         "arm64"
@@ -52,11 +52,11 @@
         "uuid": "^9.0.0"
       },
       "optionalDependencies": {
-        "@lancedb/vectordb-darwin-arm64": "0.19.0-beta.6",
+        "@lancedb/vectordb-darwin-arm64": "0.20.1-beta.2",
-        "@lancedb/vectordb-darwin-x64": "0.19.0-beta.6",
+        "@lancedb/vectordb-darwin-x64": "0.20.1-beta.2",
-        "@lancedb/vectordb-linux-arm64-gnu": "0.19.0-beta.6",
+        "@lancedb/vectordb-linux-arm64-gnu": "0.20.1-beta.2",
-        "@lancedb/vectordb-linux-x64-gnu": "0.19.0-beta.6",
+        "@lancedb/vectordb-linux-x64-gnu": "0.20.1-beta.2",
-        "@lancedb/vectordb-win32-x64-msvc": "0.19.0-beta.6"
+        "@lancedb/vectordb-win32-x64-msvc": "0.20.1-beta.2"
       },
       "peerDependencies": {
         "@apache-arrow/ts": "^14.0.2",
@@ -327,65 +327,60 @@
       }
     },
     "node_modules/@lancedb/vectordb-darwin-arm64": {
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.19.0-beta.6.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.20.1-beta.2.tgz",
-      "integrity": "sha512-fujUe3Gt1n1vgxXMDaUatZEQICh9VAmj1CJK/gQCMZo9ky/MH1TnxP0nA6hN7fkRvl28C2Ms2adlTdlnTxLSlw==",
+      "integrity": "sha512-mqi0yI+ZwBTydaDy1FRHAUZwrWS28u6tbHTe1s4uSrmERbVI6PfmoPR+NZWWAp6ZhlseSdl/+yeI4imk11rQSw==",
       "cpu": [
         "arm64"
       ],
-      "license": "Apache-2.0",
       "optional": true,
       "os": [
         "darwin"
       ]
     },
     "node_modules/@lancedb/vectordb-darwin-x64": {
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.19.0-beta.6.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.20.1-beta.2.tgz",
-      "integrity": "sha512-ZKUvPwKvnK5WfyCR3Asbm1XXXA5JWYfDVD2ovPU/mv/rqoroYEpxm7TH1OG8AQ8bvBmrCmPc0sPJP5kijd6BFg==",
+      "integrity": "sha512-m8EYYA8JZIeNsJqQsBDUMu6r31/u7FzpjonJ4Y+CjapVl6UdvI65KUkeL2dYrFao++RuIoaiqcm3e7gRgFZpXQ==",
       "cpu": [
         "x64"
       ],
-      "license": "Apache-2.0",
       "optional": true,
       "os": [
         "darwin"
       ]
     },
     "node_modules/@lancedb/vectordb-linux-arm64-gnu": {
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.19.0-beta.6.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.20.1-beta.2.tgz",
-      "integrity": "sha512-m4DuGCEhEAy+EtamSBMF1ujiVkpJD3ybF/Yp1pYYo9FTFThczAeRiyUg7diRZYfahZExKsATj62PqHXNVo8x9A==",
+      "integrity": "sha512-3Og2+bk4GlWmMO1Yg2HBfeb5zrOMLaIHD7bEqQ4+6yw4IckAaV+ke05H0tyyqmOVrOQ0LpvtXgD7pPztjm9r9A==",
       "cpu": [
         "arm64"
       ],
-      "license": "Apache-2.0",
       "optional": true,
       "os": [
         "linux"
       ]
     },
     "node_modules/@lancedb/vectordb-linux-x64-gnu": {
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.19.0-beta.6.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.20.1-beta.2.tgz",
-      "integrity": "sha512-npUR23GZJDVfkPUPtaxLuYUeqyAQ/vcp4R7RjCSdBo+hJNiQAG4TX31YAE8OKnOGskEO7XJ3BgEAxM+upiNmnA==",
+      "integrity": "sha512-mwTQyA/FBoU/FkPuvCNBZG3y83gBN+iYoejehBH2HBkLUIcmlsDgSRZ1OQ+f9ijj12EMBCA11tBUPA9zhHzyrw==",
       "cpu": [
         "x64"
       ],
-      "license": "Apache-2.0",
       "optional": true,
       "os": [
         "linux"
       ]
     },
     "node_modules/@lancedb/vectordb-win32-x64-msvc": {
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.19.0-beta.6.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.20.1-beta.2.tgz",
-      "integrity": "sha512-Ebas+phT0D7NoB1e3lMZn5h7WVyT5pPIwO1Kk1cZ93V4zaxn2BQRwjLTLxJwR9G+emQoLv659Ze0NtnFuEbXaA==",
+      "integrity": "sha512-VkjNpqhK3l3uHLLPmox+HrmKPMaZgV+qsGQWx0nfseGnSOEmXAWZWQFe0APVCQ9y0xTypQB0oH7eSOPZv2t4WQ==",
       "cpu": [
         "x64"
       ],
-      "license": "Apache-2.0",
       "optional": true,
       "os": [
         "win32"

node/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vectordb",
-  "version": "0.19.0-beta.6",
+  "version": "0.20.1-beta.2",
   "description": " Serverless, low-latency vector database for AI applications",
   "private": false,
   "main": "dist/index.js",
@@ -89,10 +89,10 @@
     }
   },
   "optionalDependencies": {
-    "@lancedb/vectordb-darwin-x64": "0.19.0-beta.6",
+    "@lancedb/vectordb-darwin-x64": "0.20.1-beta.2",
-    "@lancedb/vectordb-darwin-arm64": "0.19.0-beta.6",
+    "@lancedb/vectordb-darwin-arm64": "0.20.1-beta.2",
-    "@lancedb/vectordb-linux-x64-gnu": "0.19.0-beta.6",
+    "@lancedb/vectordb-linux-x64-gnu": "0.20.1-beta.2",
-    "@lancedb/vectordb-linux-arm64-gnu": "0.19.0-beta.6",
+    "@lancedb/vectordb-linux-arm64-gnu": "0.20.1-beta.2",
-    "@lancedb/vectordb-win32-x64-msvc": "0.19.0-beta.6"
+    "@lancedb/vectordb-win32-x64-msvc": "0.20.1-beta.2"
   }
 }

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.19.0-beta.6"
+version = "0.20.1-beta.2"
 license.workspace = true
 description.workspace = true
 repository.workspace = true
@@ -28,6 +28,10 @@ napi-derive = "2.16.4"
 lzma-sys = { version = "*", features = ["static"] }
 log.workspace = true
 
+# Workaround for build failure until we can fix it.
+aws-lc-sys = "=0.28.0"
+aws-lc-rs = "=1.13.0"
+
 [build-dependencies]
 napi-build = "2.1"
 

nodejs/__test__/arrow.test.ts

@@ -374,6 +374,71 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         expect(table2.numRows).toBe(4);
         expect(table2.schema).toEqual(schema);
       });
+
+      it("should correctly retain values in nested struct fields", async function () {
+        // Define test data with nested struct
+        const testData = [
+          {
+            id: "doc1",
+            vector: [1, 2, 3],
+            metadata: {
+              filePath: "/path/to/file1.ts",
+              startLine: 10,
+              endLine: 20,
+              text: "function test() { return true; }",
+            },
+          },
+          {
+            id: "doc2",
+            vector: [4, 5, 6],
+            metadata: {
+              filePath: "/path/to/file2.ts",
+              startLine: 30,
+              endLine: 40,
+              text: "function test2() { return false; }",
+            },
+          },
+        ];
+
+        // Create Arrow table from the data
+        const table = makeArrowTable(testData);
+
+        // Verify schema has the nested struct fields
+        const metadataField = table.schema.fields.find(
+          (f) => f.name === "metadata",
+        );
+        expect(metadataField).toBeDefined();
+        // biome-ignore lint/suspicious/noExplicitAny: accessing fields in different Arrow versions
+        const childNames = metadataField?.type.children.map((c: any) => c.name);
+        expect(childNames).toEqual([
+          "filePath",
+          "startLine",
+          "endLine",
+          "text",
+        ]);
+
+        // Convert to buffer and back (simulating storage and retrieval)
+        const buf = await fromTableToBuffer(table);
+        const retrievedTable = tableFromIPC(buf);
+
+        // Verify the retrieved table has the same structure
+        const rows = [];
+        for (let i = 0; i < retrievedTable.numRows; i++) {
+          rows.push(retrievedTable.get(i));
+        }
+
+        // Check values in the first row
+        const firstRow = rows[0];
+        expect(firstRow.id).toBe("doc1");
+        expect(firstRow.vector.toJSON()).toEqual([1, 2, 3]);
+
+        // Verify metadata values are preserved (this is where the bug is)
+        expect(firstRow.metadata).toBeDefined();
+        expect(firstRow.metadata.filePath).toBe("/path/to/file1.ts");
+        expect(firstRow.metadata.startLine).toBe(10);
+        expect(firstRow.metadata.endLine).toBe(20);
+        expect(firstRow.metadata.text).toBe("function test() { return true; }");
+      });
     });
 
     class DummyEmbedding extends EmbeddingFunction<string> {
@@ -527,14 +592,14 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         ).rejects.toThrow("column vector was missing");
       });
 
-      it("will provide a nice error if run twice", async function () {
+      it("will skip embedding application if already applied", async function () {
         const records = sampleRecords();
         const table = await convertToTable(records, dummyEmbeddingConfig);
 
         // fromTableToBuffer will try and apply the embeddings again
-        await expect(
+        // but should skip since the column already has non-null values
-          fromTableToBuffer(table, dummyEmbeddingConfig),
+        const result = await fromTableToBuffer(table, dummyEmbeddingConfig);
-        ).rejects.toThrow("already existed");
+        expect(result.byteLength).toBeGreaterThan(0);
       });
     });
 

nodejs/__test__/table.test.ts

@@ -33,7 +33,13 @@ import {
   register,
 } from "../lancedb/embedding";
 import { Index } from "../lancedb/indices";
-import { instanceOfFullTextQuery } from "../lancedb/query";
+import {
+  BooleanQuery,
+  Occur,
+  Operator,
+  instanceOfFullTextQuery,
+} from "../lancedb/query";
+import exp = require("constants");
 
 describe.each([arrow15, arrow16, arrow17, arrow18])(
   "Given a table",
@@ -71,8 +77,33 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       await expect(table.countRows()).resolves.toBe(3);
     });
 
-    it("should overwrite data if asked", async () => {
+    it("should show table stats", async () => {
       await table.add([{ id: 1 }, { id: 2 }]);
+      await table.add([{ id: 1 }]);
+      await expect(table.stats()).resolves.toEqual({
+        fragmentStats: {
+          lengths: {
+            max: 2,
+            mean: 1,
+            min: 1,
+            p25: 1,
+            p50: 2,
+            p75: 2,
+            p99: 2,
+          },
+          numFragments: 2,
+          numSmallFragments: 2,
+        },
+        numIndices: 0,
+        numRows: 3,
+        totalBytes: 24,
+      });
+    });
+
+    it("should overwrite data if asked", async () => {
+      const addRes = await table.add([{ id: 1 }, { id: 2 }]);
+      expect(addRes).toHaveProperty("version");
+      expect(addRes.version).toBe(2);
       await table.add([{ id: 1 }], { mode: "overwrite" });
       await expect(table.countRows()).resolves.toBe(1);
     });
@@ -88,7 +119,11 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       await table.add([{ id: 1 }]);
       expect(await table.countRows("id == 1")).toBe(1);
       expect(await table.countRows("id == 7")).toBe(0);
-      await table.update({ id: "7" });
+      const updateRes = await table.update({ id: "7" });
+      expect(updateRes).toHaveProperty("version");
+      expect(updateRes.version).toBe(3);
+      expect(updateRes).toHaveProperty("rowsUpdated");
+      expect(updateRes.rowsUpdated).toBe(1);
       expect(await table.countRows("id == 1")).toBe(0);
       expect(await table.countRows("id == 7")).toBe(1);
       await table.add([{ id: 2 }]);
@@ -315,11 +350,17 @@ describe("merge insert", () => {
       { a: 3, b: "y" },
       { a: 4, b: "z" },
     ];
-    await table
+    const mergeInsertRes = await table
       .mergeInsert("a")
       .whenMatchedUpdateAll()
       .whenNotMatchedInsertAll()
-      .execute(newData);
+      .execute(newData, { timeoutMs: 10_000 });
+    expect(mergeInsertRes).toHaveProperty("version");
+    expect(mergeInsertRes.version).toBe(2);
+    expect(mergeInsertRes.numInsertedRows).toBe(1);
+    expect(mergeInsertRes.numUpdatedRows).toBe(2);
+    expect(mergeInsertRes.numDeletedRows).toBe(0);
+
     const expected = [
       { a: 1, b: "a" },
       { a: 2, b: "x" },
@@ -337,10 +378,12 @@ describe("merge insert", () => {
       { a: 3, b: "y" },
       { a: 4, b: "z" },
     ];
-    await table
+    const mergeInsertRes = await table
       .mergeInsert("a")
       .whenMatchedUpdateAll({ where: "target.b = 'b'" })
       .execute(newData);
+    expect(mergeInsertRes).toHaveProperty("version");
+    expect(mergeInsertRes.version).toBe(2);
 
     const expected = [
       { a: 1, b: "a" },
@@ -425,6 +468,20 @@ describe("merge insert", () => {
     res = res.sort((a, b) => a.a - b.a);
     expect(res).toEqual(expected);
   });
+
+  test("timeout", async () => {
+    const newData = [
+      { a: 2, b: "x" },
+      { a: 4, b: "z" },
+    ];
+    await expect(
+      table
+        .mergeInsert("a")
+        .whenMatchedUpdateAll()
+        .whenNotMatchedInsertAll()
+        .execute(newData, { timeoutMs: 0 }),
+    ).rejects.toThrow("merge insert timed out");
+  });
 });
 
 describe("When creating an index", () => {
@@ -502,11 +559,46 @@ describe("When creating an index", () => {
     rst = await tbl.query().limit(2).offset(1).nearestTo(queryVec).toArrow();
     expect(rst.numRows).toBe(1);
 
+    // test nprobes
+    rst = await tbl.query().nearestTo(queryVec).limit(2).nprobes(50).toArrow();
+    expect(rst.numRows).toBe(2);
+    rst = await tbl
+      .query()
+      .nearestTo(queryVec)
+      .limit(2)
+      .minimumNprobes(15)
+      .toArrow();
+    expect(rst.numRows).toBe(2);
+    rst = await tbl
+      .query()
+      .nearestTo(queryVec)
+      .limit(2)
+      .minimumNprobes(10)
+      .maximumNprobes(20)
+      .toArrow();
+    expect(rst.numRows).toBe(2);
+
+    expect(() => tbl.query().nearestTo(queryVec).minimumNprobes(0)).toThrow(
+      "Invalid input, minimum_nprobes must be greater than 0",
+    );
+    expect(() => tbl.query().nearestTo(queryVec).maximumNprobes(5)).toThrow(
+      "Invalid input, maximum_nprobes must be greater than minimum_nprobes",
+    );
+
     await tbl.dropIndex("vec_idx");
     const indices2 = await tbl.listIndices();
     expect(indices2.length).toBe(0);
   });
 
+  it("should wait for index readiness", async () => {
+    // Create an index and then wait for it to be ready
+    await tbl.createIndex("vec");
+    const indices = await tbl.listIndices();
+    expect(indices.length).toBeGreaterThan(0);
+    const idxName = indices[0].name;
+    await expect(tbl.waitForIndex([idxName], 5)).resolves.toBeUndefined();
+  });
+
   it("should search with distance range", async () => {
     await tbl.createIndex("vec");
 
@@ -824,6 +916,7 @@ describe("When creating an index", () => {
     // Only build index over v1
     await tbl.createIndex("vec", {
       config: Index.ivfPq({ numPartitions: 2, numSubVectors: 2 }),
+      waitTimeoutSeconds: 30,
     });
 
     const rst = await tbl
@@ -990,15 +1083,19 @@ describe("schema evolution", function () {
       { id: 1n, vector: [0.1, 0.2] },
     ]);
     // Can create a non-nullable column only through addColumns at the moment.
-    await table.addColumns([
+    const addColumnsRes = await table.addColumns([
       { name: "price", valueSql: "cast(10.0 as double)" },
     ]);
+    expect(addColumnsRes).toHaveProperty("version");
+    expect(addColumnsRes.version).toBe(2);
     expect(await table.schema()).toEqual(schema);
 
-    await table.alterColumns([
+    const alterColumnsRes = await table.alterColumns([
       { path: "id", rename: "new_id" },
       { path: "price", nullable: true },
     ]);
+    expect(alterColumnsRes).toHaveProperty("version");
+    expect(alterColumnsRes.version).toBe(3);
 
     const expectedSchema = new Schema([
       new Field("new_id", new Int64(), true),
@@ -1116,7 +1213,9 @@ describe("schema evolution", function () {
     const table = await con.createTable("vectors", [
       { id: 1n, vector: [0.1, 0.2] },
     ]);
-    await table.dropColumns(["vector"]);
+    const dropColumnsRes = await table.dropColumns(["vector"]);
+    expect(dropColumnsRes).toHaveProperty("version");
+    expect(dropColumnsRes.version).toBe(2);
 
     const expectedSchema = new Schema([new Field("id", new Int64(), true)]);
     expect(await table.schema()).toEqual(expectedSchema);
@@ -1168,6 +1267,99 @@ describe("when dealing with versioning", () => {
   });
 });
 
+describe("when dealing with tags", () => {
+  let tmpDir: tmp.DirResult;
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => {
+    tmpDir.removeCallback();
+  });
+
+  it("can manage tags", async () => {
+    const conn = await connect(tmpDir.name, {
+      readConsistencyInterval: 0,
+    });
+
+    const table = await conn.createTable("my_table", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+    expect(await table.version()).toBe(1);
+
+    await table.add([{ id: 2n, vector: [0.3, 0.4] }]);
+    expect(await table.version()).toBe(2);
+
+    const tagsManager = await table.tags();
+
+    const initialTags = await tagsManager.list();
+    expect(Object.keys(initialTags).length).toBe(0);
+
+    const tag1 = "tag1";
+    await tagsManager.create(tag1, 1);
+    expect(await tagsManager.getVersion(tag1)).toBe(1);
+
+    const tagsAfterFirst = await tagsManager.list();
+    expect(Object.keys(tagsAfterFirst).length).toBe(1);
+    expect(tagsAfterFirst).toHaveProperty(tag1);
+    expect(tagsAfterFirst[tag1].version).toBe(1);
+
+    await tagsManager.create("tag2", 2);
+    expect(await tagsManager.getVersion("tag2")).toBe(2);
+
+    const tagsAfterSecond = await tagsManager.list();
+    expect(Object.keys(tagsAfterSecond).length).toBe(2);
+    expect(tagsAfterSecond).toHaveProperty(tag1);
+    expect(tagsAfterSecond[tag1].version).toBe(1);
+    expect(tagsAfterSecond).toHaveProperty("tag2");
+    expect(tagsAfterSecond["tag2"].version).toBe(2);
+
+    await table.add([{ id: 3n, vector: [0.5, 0.6] }]);
+    await tagsManager.update(tag1, 3);
+    expect(await tagsManager.getVersion(tag1)).toBe(3);
+
+    await tagsManager.delete("tag2");
+    const tagsAfterDelete = await tagsManager.list();
+    expect(Object.keys(tagsAfterDelete).length).toBe(1);
+    expect(tagsAfterDelete).toHaveProperty(tag1);
+    expect(tagsAfterDelete[tag1].version).toBe(3);
+
+    await table.add([{ id: 4n, vector: [0.7, 0.8] }]);
+    expect(await table.version()).toBe(4);
+
+    await table.checkout(tag1);
+    expect(await table.version()).toBe(3);
+
+    await table.checkoutLatest();
+    expect(await table.version()).toBe(4);
+  });
+
+  it("can checkout and restore tags", async () => {
+    const conn = await connect(tmpDir.name, {
+      readConsistencyInterval: 0,
+    });
+
+    const table = await conn.createTable("my_table", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+    expect(await table.version()).toBe(1);
+    expect(await table.countRows()).toBe(1);
+    const tagsManager = await table.tags();
+    const tag1 = "tag1";
+    await tagsManager.create(tag1, 1);
+    await table.add([{ id: 2n, vector: [0.3, 0.4] }]);
+    const tag2 = "tag2";
+    await tagsManager.create(tag2, 2);
+    expect(await table.version()).toBe(2);
+    await table.checkout(tag1);
+    expect(await table.version()).toBe(1);
+    await table.restore();
+    expect(await table.version()).toBe(3);
+    expect(await table.countRows()).toBe(1);
+    await table.add([{ id: 3n, vector: [0.5, 0.6] }]);
+    expect(await table.countRows()).toBe(2);
+  });
+});
+
 describe("when optimizing a dataset", () => {
   let tmpDir: tmp.DirResult;
   let table: Table;
@@ -1312,7 +1504,7 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(results2[0].text).toBe(data[1].text);
     });
 
-    test("full text index on list", async () => {
+    test("prewarm full text search index", async () => {
       const db = await connect(tmpDir.name);
       const data = [
         { text: ["lance database", "the", "search"], vector: [0.1, 0.2, 0.3] },
@@ -1326,6 +1518,30 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         config: Index.fts(),
       });
 
+      // For the moment, we just confirm we can call prewarmIndex without error
+      // and still search it afterwards
+      await table.prewarmIndex("text_idx");
+
+      const results = await table.search("lance").toArray();
+      expect(results.length).toBe(3);
+    });
+
+    test("full text index on list", async () => {
+      const db = await connect(tmpDir.name);
+      const data = [
+        { text: ["lance database", "the", "search"], vector: [0.1, 0.2, 0.3] },
+        { text: ["lance database"], vector: [0.4, 0.5, 0.6] },
+        { text: ["lance", "search"], vector: [0.7, 0.8, 0.9] },
+        { text: ["database", "search"], vector: [1.0, 1.1, 1.2] },
+        { text: ["unrelated", "doc"], vector: [1.3, 1.4, 1.5] },
+      ];
+      const table = await db.createTable("test", data);
+      await table.createIndex("text", {
+        config: Index.fts({
+          withPosition: true,
+        }),
+      });
+
       const results = await table.search("lance").toArray();
       expect(results.length).toBe(3);
 
@@ -1346,6 +1562,18 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
 
       const results = await table.search("hello").toArray();
       expect(results[0].text).toBe(data[0].text);
+
+      const results2 = await table
+        .search(new MatchQuery("hello world", "text"))
+        .toArray();
+      expect(results2.length).toBe(2);
+
+      const results3 = await table
+        .search(
+          new MatchQuery("hello world", "text", { operator: Operator.And }),
+        )
+        .toArray();
+      expect(results3.length).toBe(1);
     });
 
     test("full text search without lowercase", async () => {
@@ -1376,7 +1604,9 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       ];
       const table = await db.createTable("test", data);
       await table.createIndex("text", {
-        config: Index.fts(),
+        config: Index.fts({
+          withPosition: true,
+        }),
       });
 
       const results = await table.search("world").toArray();
@@ -1420,6 +1650,60 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(resultSet.has("fob")).toBe(true);
       expect(resultSet.has("fo")).toBe(true);
       expect(resultSet.has("food")).toBe(true);
+
+      const prefixResults = await table
+        .search(
+          new MatchQuery("foo", "text", { fuzziness: 3, prefixLength: 3 }),
+        )
+        .toArray();
+      expect(prefixResults.length).toBe(2);
+      const resultSet2 = new Set(prefixResults.map((r) => r.text));
+      expect(resultSet2.has("foo")).toBe(true);
+      expect(resultSet2.has("food")).toBe(true);
+    });
+
+    test("full text search boolean query", async () => {
+      const db = await connect(tmpDir.name);
+      const data = [
+        { text: "The cat and dog are playing" },
+        { text: "The cat is sleeping" },
+        { text: "The dog is barking" },
+        { text: "The dog chases the cat" },
+      ];
+      const table = await db.createTable("test", data);
+      await table.createIndex("text", {
+        config: Index.fts({ withPosition: false }),
+      });
+
+      const shouldResults = await table
+        .search(
+          new BooleanQuery([
+            [Occur.Should, new MatchQuery("cat", "text")],
+            [Occur.Should, new MatchQuery("dog", "text")],
+          ]),
+        )
+        .toArray();
+      expect(shouldResults.length).toBe(4);
+
+      const mustResults = await table
+        .search(
+          new BooleanQuery([
+            [Occur.Must, new MatchQuery("cat", "text")],
+            [Occur.Must, new MatchQuery("dog", "text")],
+          ]),
+        )
+        .toArray();
+      expect(mustResults.length).toBe(2);
+
+      const mustNotResults = await table
+        .search(
+          new BooleanQuery([
+            [Occur.Must, new MatchQuery("cat", "text")],
+            [Occur.MustNot, new MatchQuery("dog", "text")],
+          ]),
+        )
+        .toArray();
+      expect(mustNotResults.length).toBe(1);
     });
 
     test.each([

nodejs/lancedb/arrow.ts

@@ -417,7 +417,9 @@ function inferSchema(
         } else {
           const inferredType = inferType(value, path, opts);
           if (inferredType === undefined) {
-            throw new Error(`Failed to infer data type for field ${path.join(".")} at row ${rowI}. \
+            throw new Error(`Failed to infer data type for field ${path.join(
+              ".",
+            )} at row ${rowI}. \
                              Consider providing an explicit schema.`);
           }
           pathTree.set(path, inferredType);
@@ -639,8 +641,9 @@ function transposeData(
 ): Vector {
   if (field.type instanceof Struct) {
     const childFields = field.type.children;
+    const fullPath = [...path, field.name];
     const childVectors = childFields.map((child) => {
-      return transposeData(data, child, [...path, child.name]);
+      return transposeData(data, child, fullPath);
     });
     const structData = makeData({
       type: field.type,
@@ -652,7 +655,14 @@ function transposeData(
     const values = data.map((datum) => {
       let current: unknown = datum;
       for (const key of valuesPath) {
-        if (isObject(current) && Object.hasOwn(current, key)) {
+        if (current == null) {
+          return null;
+        }
+
+        if (
+          isObject(current) &&
+          (Object.hasOwn(current, key) || key in current)
+        ) {
           current = current[key];
         } else {
           return null;
@@ -791,11 +801,17 @@ async function applyEmbeddingsFromMetadata(
         `Cannot apply embedding function because the source column '${functionEntry.sourceColumn}' was not present in the data`,
       );
     }
+
+    // Check if destination column exists and handle accordingly
     if (columns[destColumn] !== undefined) {
-      throw new Error(
+      const existingColumn = columns[destColumn];
-        `Attempt to apply embeddings to table failed because column ${destColumn} already existed`,
+      // If the column exists but is all null, we can fill it with embeddings
-      );
+      if (existingColumn.nullCount !== existingColumn.length) {
+        // Column has non-null values, skip embedding application
+        continue;
+      }
     }
+
     if (table.batches.length > 1) {
       throw new Error(
         "Internal error: `makeArrowTable` unexpectedly created a table with more than one batch",
@@ -895,11 +911,23 @@ async function applyEmbeddings<T>(
       );
     }
   } else {
+    // Check if destination column exists and handle accordingly
     if (Object.prototype.hasOwnProperty.call(newColumns, destColumn)) {
-      throw new Error(
+      const existingColumn = newColumns[destColumn];
-        `Attempt to apply embeddings to table failed because column ${destColumn} already existed`,
+      // If the column exists but is all null, we can fill it with embeddings
-      );
+      if (existingColumn.nullCount !== existingColumn.length) {
+        // Column has non-null values, skip embedding application and return table as-is
+        let newTable = new ArrowTable(newColumns);
+        if (schema != null) {
+          newTable = alignTable(newTable, schema as Schema);
+        }
+        return new ArrowTable(
+          new Schema(newTable.schema.fields, schemaMetadata),
+          newTable.batches,
+        );
+      }
     }
+
     if (table.batches.length > 1) {
       throw new Error(
         "Internal error: `makeArrowTable` unexpectedly created a table with more than one batch",

nodejs/lancedb/index.ts

@@ -23,6 +23,18 @@ export {
   OptimizeStats,
   CompactionStats,
   RemovalStats,
+  TableStatistics,
+  FragmentStatistics,
+  FragmentSummaryStats,
+  Tags,
+  TagContents,
+  MergeResult,
+  AddResult,
+  AddColumnsResult,
+  AlterColumnsResult,
+  DeleteResult,
+  DropColumnsResult,
+  UpdateResult,
 } from "./native.js";
 
 export {
@@ -52,7 +64,10 @@ export {
   PhraseQuery,
   BoostQuery,
   MultiMatchQuery,
+  BooleanQuery,
   FullTextQueryType,
+  Operator,
+  Occur,
 } from "./query";
 
 export {
@@ -74,7 +89,7 @@ export {
   ColumnAlteration,
 } from "./table";
 
-export { MergeInsertBuilder } from "./merge";
+export { MergeInsertBuilder, WriteExecutionOptions } from "./merge";
 
 export * as embedding from "./embedding";
 export * as rerankers from "./rerankers";

nodejs/lancedb/indices.ts

@@ -681,4 +681,6 @@ export interface IndexOptions {
    * The default is true
    */
   replace?: boolean;
+
+  waitTimeoutSeconds?: number;
 }

nodejs/lancedb/merge.ts

@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 import { Data, Schema, fromDataToBuffer } from "./arrow";
-import { NativeMergeInsertBuilder } from "./native";
+import { MergeResult, NativeMergeInsertBuilder } from "./native";
 
 /** A builder used to create and run a merge insert operation */
 export class MergeInsertBuilder {
@@ -73,9 +73,12 @@ export class MergeInsertBuilder {
   /**
    * Executes the merge insert operation
    *
-   * Nothing is returned but the `Table` is updated
+   * @returns {Promise<MergeResult>} the merge result
    */
-  async execute(data: Data): Promise<void> {
+  async execute(
+    data: Data,
+    execOptions?: Partial<WriteExecutionOptions>,
+  ): Promise<MergeResult> {
     let schema: Schema;
     if (this.#schema instanceof Promise) {
       schema = await this.#schema;
@@ -83,7 +86,28 @@ export class MergeInsertBuilder {
     } else {
       schema = this.#schema;
     }
+
+    if (execOptions?.timeoutMs !== undefined) {
+      this.#native.setTimeout(execOptions.timeoutMs);
+    }
+
     const buffer = await fromDataToBuffer(data, undefined, schema);
-    await this.#native.execute(buffer);
+    return await this.#native.execute(buffer);
   }
 }
+
+export interface WriteExecutionOptions {
+  /**
+   * Maximum time to run the operation before cancelling it.
+   *
+   * By default, there is a 30-second timeout that is only enforced after the
+   * first attempt. This is to prevent spending too long retrying to resolve
+   * conflicts. For example, if a write attempt takes 20 seconds and fails,
+   * the second attempt will be cancelled after 10 seconds, hitting the
+   * 30-second timeout. However, a write that takes one hour and succeeds on the
+   * first attempt will not be cancelled.
+   *
+   * When this is set, the timeout is enforced on all attempts, including the first.
+   */
+  timeoutMs?: number;
+}

nodejs/lancedb/query.ts

@@ -448,6 +448,10 @@ export class VectorQuery extends QueryBase<NativeVectorQuery> {
    * For best results we recommend tuning this parameter with a benchmark against
    * your actual data to find the smallest possible value that will still give
    * you the desired recall.
+   *
+   * For more fine grained control over behavior when you have a very narrow filter
+   * you can use `minimumNprobes` and `maximumNprobes`.  This method sets both
+   * the minimum and maximum to the same value.
    */
   nprobes(nprobes: number): VectorQuery {
     super.doCall((inner) => inner.nprobes(nprobes));
@@ -455,6 +459,33 @@ export class VectorQuery extends QueryBase<NativeVectorQuery> {
     return this;
   }
 
+  /**
+   * Set the minimum number of probes used.
+   *
+   * This controls the minimum number of partitions that will be searched.  This
+   * parameter will impact every query against a vector index, regardless of the
+   * filter.  See `nprobes` for more details.  Higher values will increase recall
+   * but will also increase latency.
+   */
+  minimumNprobes(minimumNprobes: number): VectorQuery {
+    super.doCall((inner) => inner.minimumNprobes(minimumNprobes));
+    return this;
+  }
+
+  /**
+   * Set the maximum number of probes used.
+   *
+   * This controls the maximum number of partitions that will be searched.  If this
+   * number is greater than minimumNprobes then the excess partitions will _only_ be
+   * searched if we have not found enough results.  This can be useful when there is
+   * a narrow filter to allow these queries to spend more time searching and avoid
+   * potential false negatives.
+   */
+  maximumNprobes(maximumNprobes: number): VectorQuery {
+    super.doCall((inner) => inner.maximumNprobes(maximumNprobes));
+    return this;
+  }
+
   /*
    * Set the distance range to use
    *
@@ -762,6 +793,31 @@ export enum FullTextQueryType {
   MatchPhrase = "match_phrase",
   Boost = "boost",
   MultiMatch = "multi_match",
+  Boolean = "boolean",
+}
+
+/**
+ * Enum representing the logical operators used in full-text queries.
+ *
+ * - `And`: All terms must match.
+ * - `Or`: At least one term must match.
+ */
+export enum Operator {
+  And = "AND",
+  Or = "OR",
+}
+
+/**
+ * Enum representing the occurrence of terms in full-text queries.
+ *
+ * - `Must`: The term must be present in the document.
+ * - `Should`: The term should contribute to the document score, but is not required.
+ * - `MustNot`: The term must not be present in the document.
+ */
+export enum Occur {
+  Should = "SHOULD",
+  Must = "MUST",
+  MustNot = "MUST_NOT",
 }
 
 /**
@@ -791,6 +847,7 @@ export function instanceOfFullTextQuery(obj: any): obj is FullTextQuery {
 export class MatchQuery implements FullTextQuery {
   /** @ignore */
   public readonly inner: JsFullTextQuery;
+
   /**
    * Creates an instance of MatchQuery.
    *
@@ -800,6 +857,8 @@ export class MatchQuery implements FullTextQuery {
    *   - `boost`: The boost factor for the query (default is 1.0).
    *   - `fuzziness`: The fuzziness level for the query (default is 0).
    *   - `maxExpansions`: The maximum number of terms to consider for fuzzy matching (default is 50).
+   *   - `operator`: The logical operator to use for combining terms in the query (default is "OR").
+   *   - `prefixLength`: The number of beginning characters being unchanged for fuzzy matching.
    */
   constructor(
     query: string,
@@ -808,6 +867,8 @@ export class MatchQuery implements FullTextQuery {
       boost?: number;
       fuzziness?: number;
       maxExpansions?: number;
+      operator?: Operator;
+      prefixLength?: number;
     },
   ) {
     let fuzziness = options?.fuzziness;
@@ -820,6 +881,8 @@ export class MatchQuery implements FullTextQuery {
       options?.boost ?? 1.0,
       fuzziness,
       options?.maxExpansions ?? 50,
+      options?.operator ?? Operator.Or,
+      options?.prefixLength ?? 0,
     );
   }
 
@@ -836,9 +899,11 @@ export class PhraseQuery implements FullTextQuery {
    *
    * @param query - The phrase to search for in the specified column.
    * @param column - The name of the column to search within.
+   * @param options - Optional parameters for the phrase query.
+   *   - `slop`: The maximum number of intervening unmatched positions allowed between words in the phrase (default is 0).
    */
-  constructor(query: string, column: string) {
+  constructor(query: string, column: string, options?: { slop?: number }) {
-    this.inner = JsFullTextQuery.phraseQuery(query, column);
+    this.inner = JsFullTextQuery.phraseQuery(query, column, options?.slop ?? 0);
   }
 
   queryType(): FullTextQueryType {
@@ -889,18 +954,21 @@ export class MultiMatchQuery implements FullTextQuery {
    * @param columns - An array of column names to search within.
    * @param options - Optional parameters for the multi-match query.
    *  - `boosts`: An array of boost factors for each column (default is 1.0 for all).
+   *  - `operator`: The logical operator to use for combining terms in the query (default is "OR").
    */
   constructor(
     query: string,
     columns: string[],
     options?: {
       boosts?: number[];
+      operator?: Operator;
     },
   ) {
     this.inner = JsFullTextQuery.multiMatchQuery(
       query,
       columns,
       options?.boosts,
+      options?.operator ?? Operator.Or,
     );
   }
 
@@ -908,3 +976,23 @@ export class MultiMatchQuery implements FullTextQuery {
     return FullTextQueryType.MultiMatch;
   }
 }
+
+export class BooleanQuery implements FullTextQuery {
+  /** @ignore */
+  public readonly inner: JsFullTextQuery;
+  /**
+   * Creates an instance of BooleanQuery.
+   *
+   * @param queries - An array of (Occur, FullTextQuery objects) to combine.
+   * Occur specifies whether the query must match, or should match.
+   */
+  constructor(queries: [Occur, FullTextQuery][]) {
+    this.inner = JsFullTextQuery.booleanQuery(
+      queries.map(([occur, query]) => [occur, query.inner]),
+    );
+  }
+
+  queryType(): FullTextQueryType {
+    return FullTextQueryType.Boolean;
+  }
+}

nodejs/lancedb/table.ts

@@ -16,10 +16,18 @@ import { EmbeddingFunctionConfig, getRegistry } from "./embedding/registry";
 import { IndexOptions } from "./indices";
 import { MergeInsertBuilder } from "./merge";
 import {
+  AddColumnsResult,
   AddColumnsSql,
+  AddResult,
+  AlterColumnsResult,
+  DeleteResult,
+  DropColumnsResult,
   IndexConfig,
   IndexStatistics,
   OptimizeStats,
+  TableStatistics,
+  Tags,
+  UpdateResult,
   Table as _NativeTable,
 } from "./native";
 import {
@@ -124,12 +132,19 @@ export abstract class Table {
   /**
    * Insert records into this Table.
    * @param {Data} data Records to be inserted into the Table
+   * @returns {Promise<AddResult>} A promise that resolves to an object
+   * containing the new version number of the table
    */
-  abstract add(data: Data, options?: Partial<AddDataOptions>): Promise<void>;
+  abstract add(
+    data: Data,
+    options?: Partial<AddDataOptions>,
+  ): Promise<AddResult>;
   /**
    * Update existing records in the Table
    * @param opts.values The values to update. The keys are the column names and the values
    * are the values to set.
+   * @returns {Promise<UpdateResult>} A promise that resolves to an object containing
+   * the number of rows updated and the new version number
    * @example
    * ```ts
    * table.update({where:"x = 2", values:{"vector": [10, 10]}})
@@ -139,11 +154,13 @@ export abstract class Table {
     opts: {
       values: Map<string, IntoSql> | Record<string, IntoSql>;
     } & Partial<UpdateOptions>,
-  ): Promise<void>;
+  ): Promise<UpdateResult>;
   /**
    * Update existing records in the Table
    * @param opts.valuesSql The values to update. The keys are the column names and the values
    * are the values to set. The values are SQL expressions.
+   * @returns {Promise<UpdateResult>} A promise that resolves to an object containing
+   * the number of rows updated and the new version number
    * @example
    * ```ts
    * table.update({where:"x = 2", valuesSql:{"x": "x + 1"}})
@@ -153,7 +170,7 @@ export abstract class Table {
     opts: {
       valuesSql: Map<string, string> | Record<string, string>;
     } & Partial<UpdateOptions>,
-  ): Promise<void>;
+  ): Promise<UpdateResult>;
   /**
    * Update existing records in the Table
    *
@@ -171,6 +188,8 @@ export abstract class Table {
    * repeatedly calilng this method.
    * @param {Map<string, string> | Record<string, string>} updates - the
    * columns to update
+   * @returns {Promise<UpdateResult>} A promise that resolves to an object
+   * containing the number of rows updated and the new version number
    *
    * Keys in the map should specify the name of the column to update.
    * Values in the map provide the new value of the column.  These can
@@ -182,12 +201,16 @@ export abstract class Table {
   abstract update(
     updates: Map<string, string> | Record<string, string>,
     options?: Partial<UpdateOptions>,
-  ): Promise<void>;
+  ): Promise<UpdateResult>;
 
   /** Count the total number of rows in the dataset. */
   abstract countRows(filter?: string): Promise<number>;
-  /** Delete the rows that satisfy the predicate. */
+  /**
-  abstract delete(predicate: string): Promise<void>;
+   * Delete the rows that satisfy the predicate.
+   * @returns {Promise<DeleteResult>} A promise that resolves to an object
+   * containing the new version number of the table
+   */
+  abstract delete(predicate: string): Promise<DeleteResult>;
   /**
    * Create an index to speed up queries.
    *
@@ -235,6 +258,30 @@ export abstract class Table {
    */
   abstract dropIndex(name: string): Promise<void>;
 
+  /**
+   * Prewarm an index in the table.
+   *
+   * @param name The name of the index.
+   *
+   * This will load the index into memory.  This may reduce the cold-start time for
+   * future queries.  If the index does not fit in the cache then this call may be
+   * wasteful.
+   */
+  abstract prewarmIndex(name: string): Promise<void>;
+
+  /**
+   * Waits for asynchronous indexing to complete on the table.
+   *
+   * @param indexNames The name of the indices to wait for
+   * @param timeoutSeconds The number of seconds to wait before timing out
+   *
+   * This will raise an error if the indices are not created and fully indexed within the timeout.
+   */
+  abstract waitForIndex(
+    indexNames: string[],
+    timeoutSeconds: number,
+  ): Promise<void>;
+
   /**
    * Create a {@link Query} Builder.
    *
@@ -317,15 +364,23 @@ export abstract class Table {
    * the SQL expression to use to calculate the value of the new column. These
    * expressions will be evaluated for each row in the table, and can
    * reference existing columns in the table.
+   * @returns {Promise<AddColumnsResult>} A promise that resolves to an object
+   * containing the new version number of the table after adding the columns.
    */
-  abstract addColumns(newColumnTransforms: AddColumnsSql[]): Promise<void>;
+  abstract addColumns(
+    newColumnTransforms: AddColumnsSql[],
+  ): Promise<AddColumnsResult>;
 
   /**
    * Alter the name or nullability of columns.
    * @param {ColumnAlteration[]} columnAlterations One or more alterations to
    * apply to columns.
+   * @returns {Promise<AlterColumnsResult>} A promise that resolves to an object
+   * containing the new version number of the table after altering the columns.
    */
-  abstract alterColumns(columnAlterations: ColumnAlteration[]): Promise<void>;
+  abstract alterColumns(
+    columnAlterations: ColumnAlteration[],
+  ): Promise<AlterColumnsResult>;
   /**
    * Drop one or more columns from the dataset
    *
@@ -336,8 +391,10 @@ export abstract class Table {
    * @param {string[]} columnNames The names of the columns to drop. These can
    * be nested column references (e.g. "a.b.c") or top-level column names
    * (e.g. "a").
+   * @returns {Promise<DropColumnsResult>} A promise that resolves to an object
+   * containing the new version number of the table after dropping the columns.
    */
-  abstract dropColumns(columnNames: string[]): Promise<void>;
+  abstract dropColumns(columnNames: string[]): Promise<DropColumnsResult>;
   /** Retrieve the version of the table */
 
   abstract version(): Promise<number>;
@@ -350,7 +407,7 @@ export abstract class Table {
    *
    * Calling this method will set the table into time-travel mode. If you
    * wish to return to standard mode, call `checkoutLatest`.
-   * @param {number} version The version to checkout
+   * @param {number | string} version The version to checkout, could be version number or tag
    * @example
    * ```typescript
    * import * as lancedb from "@lancedb/lancedb"
@@ -366,7 +423,8 @@ export abstract class Table {
    * console.log(await table.version()); // 2
    * ```
    */
-  abstract checkout(version: number): Promise<void>;
+  abstract checkout(version: number | string): Promise<void>;
+
   /**
    * Checkout the latest version of the table. _This is an in-place operation._
    *
@@ -380,6 +438,23 @@ export abstract class Table {
    */
   abstract listVersions(): Promise<Version[]>;
 
+  /**
+   * Get a tags manager for this table.
+   *
+   * Tags allow you to label specific versions of a table with a human-readable name.
+   * The returned tags manager can be used to list, create, update, or delete tags.
+   *
+   * @returns {Tags} A tags manager for this table
+   * @example
+   * ```typescript
+   * const tagsManager = await table.tags();
+   * await tagsManager.create("v1", 1);
+   * const tags = await tagsManager.list();
+   * console.log(tags); // { "v1": { version: 1, manifestSize: ... } }
+   * ```
+   */
+  abstract tags(): Promise<Tags>;
+
   /**
    * Restore the table to the currently checked out version
    *
@@ -439,6 +514,13 @@ export abstract class Table {
    * Use {@link Table.listIndices} to find the names of the indices.
    */
   abstract indexStats(name: string): Promise<IndexStatistics | undefined>;
+
+  /** Returns table and fragment statistics
+   *
+   * @returns {TableStatistics} The table and fragment statistics
+   *
+   */
+  abstract stats(): Promise<TableStatistics>;
 }
 
 export class LocalTable extends Table {
@@ -478,12 +560,12 @@ export class LocalTable extends Table {
     return tbl.schema;
   }
 
-  async add(data: Data, options?: Partial<AddDataOptions>): Promise<void> {
+  async add(data: Data, options?: Partial<AddDataOptions>): Promise<AddResult> {
     const mode = options?.mode ?? "append";
     const schema = await this.schema();
 
     const buffer = await fromDataToBuffer(data, undefined, schema);
-    await this.inner.add(buffer, mode);
+    return await this.inner.add(buffer, mode);
   }
 
   async update(
@@ -496,7 +578,7 @@ export class LocalTable extends Table {
           valuesSql: Map<string, string> | Record<string, string>;
         } & Partial<UpdateOptions>),
     options?: Partial<UpdateOptions>,
-  ) {
+  ): Promise<UpdateResult> {
     const isValues =
       "values" in optsOrUpdates && typeof optsOrUpdates.values !== "string";
     const isValuesSql =
@@ -543,28 +625,44 @@ export class LocalTable extends Table {
         columns = Object.entries(optsOrUpdates as Record<string, string>);
         predicate = options?.where;
     }
-    await this.inner.update(predicate, columns);
+    return await this.inner.update(predicate, columns);
   }
 
   async countRows(filter?: string): Promise<number> {
     return await this.inner.countRows(filter);
   }
 
-  async delete(predicate: string): Promise<void> {
+  async delete(predicate: string): Promise<DeleteResult> {
-    await this.inner.delete(predicate);
+    return await this.inner.delete(predicate);
   }
 
   async createIndex(column: string, options?: Partial<IndexOptions>) {
     // Bit of a hack to get around the fact that TS has no package-scope.
     // biome-ignore lint/suspicious/noExplicitAny: skip
     const nativeIndex = (options?.config as any)?.inner;
-    await this.inner.createIndex(nativeIndex, column, options?.replace);
+    await this.inner.createIndex(
+      nativeIndex,
+      column,
+      options?.replace,
+      options?.waitTimeoutSeconds,
+    );
   }
 
   async dropIndex(name: string): Promise<void> {
     await this.inner.dropIndex(name);
   }
 
+  async prewarmIndex(name: string): Promise<void> {
+    await this.inner.prewarmIndex(name);
+  }
+
+  async waitForIndex(
+    indexNames: string[],
+    timeoutSeconds: number,
+  ): Promise<void> {
+    await this.inner.waitForIndex(indexNames, timeoutSeconds);
+  }
+
   query(): Query {
     return new Query(this.inner);
   }
@@ -623,11 +721,15 @@ export class LocalTable extends Table {
 
   // TODO: Support BatchUDF
 
-  async addColumns(newColumnTransforms: AddColumnsSql[]): Promise<void> {
+  async addColumns(
-    await this.inner.addColumns(newColumnTransforms);
+    newColumnTransforms: AddColumnsSql[],
+  ): Promise<AddColumnsResult> {
+    return await this.inner.addColumns(newColumnTransforms);
   }
 
-  async alterColumns(columnAlterations: ColumnAlteration[]): Promise<void> {
+  async alterColumns(
+    columnAlterations: ColumnAlteration[],
+  ): Promise<AlterColumnsResult> {
     const processedAlterations = columnAlterations.map((alteration) => {
       if (typeof alteration.dataType === "string") {
         return {
@@ -648,19 +750,22 @@ export class LocalTable extends Table {
       }
     });
 
-    await this.inner.alterColumns(processedAlterations);
+    return await this.inner.alterColumns(processedAlterations);
   }
 
-  async dropColumns(columnNames: string[]): Promise<void> {
+  async dropColumns(columnNames: string[]): Promise<DropColumnsResult> {
-    await this.inner.dropColumns(columnNames);
+    return await this.inner.dropColumns(columnNames);
   }
 
   async version(): Promise<number> {
     return await this.inner.version();
   }
 
-  async checkout(version: number): Promise<void> {
+  async checkout(version: number | string): Promise<void> {
-    await this.inner.checkout(version);
+    if (typeof version === "string") {
+      return this.inner.checkoutTag(version);
+    }
+    return this.inner.checkout(version);
   }
 
   async checkoutLatest(): Promise<void> {
@@ -679,6 +784,10 @@ export class LocalTable extends Table {
     await this.inner.restore();
   }
 
+  async tags(): Promise<Tags> {
+    return await this.inner.tags();
+  }
+
   async optimize(options?: Partial<OptimizeOptions>): Promise<OptimizeStats> {
     let cleanupOlderThanMs;
     if (
@@ -709,6 +818,11 @@ export class LocalTable extends Table {
     }
     return stats;
   }
+
+  async stats(): Promise<TableStatistics> {
+    return await this.inner.stats();
+  }
+
   mergeInsert(on: string | string[]): MergeInsertBuilder {
     on = Array.isArray(on) ? on : [on];
     return new MergeInsertBuilder(this.inner.mergeInsert(on), this.schema());

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

nodejs/npm/darwin-x64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-x64",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["darwin"],
 	"cpu": ["x64"],
 	"main": "lancedb.darwin-x64.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

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

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.19.0-beta.6",
+  "version": "0.20.1-beta.2",
   "os": [
     "win32"
   ],

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.19.0-beta.6",
+	"version": "0.20.1-beta.2",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@lancedb/lancedb",
-  "version": "0.19.0-beta.6",
+  "version": "0.20.1-beta.2",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@lancedb/lancedb",
-      "version": "0.19.0-beta.6",
+      "version": "0.20.1-beta.2",
       "cpu": [
         "x64",
         "arm64"

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.19.0-beta.6",
+  "version": "0.20.1-beta.2",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",

nodejs/src/index.rs

@@ -125,32 +125,30 @@ impl Index {
         ascii_folding: Option<bool>,
     ) -> Self {
         let mut opts = FtsIndexBuilder::default();
-        let mut tokenizer_configs = opts.tokenizer_configs.clone();
         if let Some(with_position) = with_position {
             opts = opts.with_position(with_position);
         }
         if let Some(base_tokenizer) = base_tokenizer {
-            tokenizer_configs = tokenizer_configs.base_tokenizer(base_tokenizer);
+            opts = opts.base_tokenizer(base_tokenizer);
         }
         if let Some(language) = language {
-            tokenizer_configs = tokenizer_configs.language(&language).unwrap();
+            opts = opts.language(&language).unwrap();
         }
         if let Some(max_token_length) = max_token_length {
-            tokenizer_configs = tokenizer_configs.max_token_length(Some(max_token_length as usize));
+            opts = opts.max_token_length(Some(max_token_length as usize));
         }
         if let Some(lower_case) = lower_case {
-            tokenizer_configs = tokenizer_configs.lower_case(lower_case);
+            opts = opts.lower_case(lower_case);
         }
         if let Some(stem) = stem {
-            tokenizer_configs = tokenizer_configs.stem(stem);
+            opts = opts.stem(stem);
         }
         if let Some(remove_stop_words) = remove_stop_words {
-            tokenizer_configs = tokenizer_configs.remove_stop_words(remove_stop_words);
+            opts = opts.remove_stop_words(remove_stop_words);
         }
         if let Some(ascii_folding) = ascii_folding {
-            tokenizer_configs = tokenizer_configs.ascii_folding(ascii_folding);
+            opts = opts.ascii_folding(ascii_folding);
         }
-        opts.tokenizer_configs = tokenizer_configs;
 
         Self {
             inner: Mutex::new(Some(LanceDbIndex::FTS(opts))),

nodejs/src/merge.rs

@@ -1,11 +1,13 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
+use std::time::Duration;
+
 use lancedb::{arrow::IntoArrow, ipc::ipc_file_to_batches, table::merge::MergeInsertBuilder};
 use napi::bindgen_prelude::*;
 use napi_derive::napi;
 
-use crate::error::convert_error;
+use crate::{error::convert_error, table::MergeResult};
 
 #[napi]
 #[derive(Clone)]
@@ -36,8 +38,13 @@ impl NativeMergeInsertBuilder {
         this
     }
 
+    #[napi]
+    pub fn set_timeout(&mut self, timeout: u32) {
+        self.inner.timeout(Duration::from_millis(timeout as u64));
+    }
+
     #[napi(catch_unwind)]
-    pub async fn execute(&self, buf: Buffer) -> napi::Result<()> {
+    pub async fn execute(&self, buf: Buffer) -> napi::Result<MergeResult> {
         let data = ipc_file_to_batches(buf.to_vec())
             .and_then(IntoArrow::into_arrow)
             .map_err(|e| {
@@ -46,12 +53,13 @@ impl NativeMergeInsertBuilder {
 
         let this = self.clone();
 
-        this.inner.execute(data).await.map_err(|e| {
+        let res = this.inner.execute(data).await.map_err(|e| {
             napi::Error::from_reason(format!(
                 "Failed to execute merge insert: {}",
                 convert_error(&e)
             ))
-        })
+        })?;
+        Ok(res.into())
     }
 }
 

nodejs/src/query.rs

@@ -4,7 +4,8 @@
 use std::sync::Arc;
 
 use lancedb::index::scalar::{
-    BoostQuery, FtsQuery, FullTextSearchQuery, MatchQuery, MultiMatchQuery, PhraseQuery,
+    BooleanQuery, BoostQuery, FtsQuery, FullTextSearchQuery, MatchQuery, MultiMatchQuery, Occur,
+    Operator, PhraseQuery,
 };
 use lancedb::query::ExecutableQuery;
 use lancedb::query::Query as LanceDbQuery;
@@ -177,6 +178,31 @@ impl VectorQuery {
         self.inner = self.inner.clone().nprobes(nprobe as usize);
     }
 
+    #[napi]
+    pub fn minimum_nprobes(&mut self, minimum_nprobe: u32) -> napi::Result<()> {
+        self.inner = self
+            .inner
+            .clone()
+            .minimum_nprobes(minimum_nprobe as usize)
+            .default_error()?;
+        Ok(())
+    }
+
+    #[napi]
+    pub fn maximum_nprobes(&mut self, maximum_nprobes: u32) -> napi::Result<()> {
+        let maximum_nprobes = if maximum_nprobes == 0 {
+            None
+        } else {
+            Some(maximum_nprobes as usize)
+        };
+        self.inner = self
+            .inner
+            .clone()
+            .maximum_nprobes(maximum_nprobes)
+            .default_error()?;
+        Ok(())
+    }
+
     #[napi]
     pub fn distance_range(&mut self, lower_bound: Option<f64>, upper_bound: Option<f64>) {
         // napi doesn't support f32, so we have to convert to f32
@@ -308,6 +334,8 @@ impl JsFullTextQuery {
         boost: f64,
         fuzziness: Option<u32>,
         max_expansions: u32,
+        operator: String,
+        prefix_length: u32,
     ) -> napi::Result<Self> {
         Ok(Self {
             inner: MatchQuery::new(query)
@@ -315,18 +343,28 @@ impl JsFullTextQuery {
                 .with_boost(boost as f32)
                 .with_fuzziness(fuzziness)
                 .with_max_expansions(max_expansions as usize)
+                .with_operator(
+                    Operator::try_from(operator.as_str()).map_err(|e| {
+                        napi::Error::from_reason(format!("Invalid operator: {}", e))
+                    })?,
+                )
+                .with_prefix_length(prefix_length)
                 .into(),
         })
     }
 
     #[napi(factory)]
-    pub fn phrase_query(query: String, column: String) -> napi::Result<Self> {
+    pub fn phrase_query(query: String, column: String, slop: u32) -> napi::Result<Self> {
         Ok(Self {
-            inner: PhraseQuery::new(query).with_column(Some(column)).into(),
+            inner: PhraseQuery::new(query)
+                .with_column(Some(column))
+                .with_slop(slop)
+                .into(),
         })
     }
 
     #[napi(factory)]
+    #[allow(clippy::use_self)] // NAPI doesn't allow Self here but clippy reports it
     pub fn boost_query(
         positive: &JsFullTextQuery,
         negative: &JsFullTextQuery,
@@ -347,20 +385,48 @@ impl JsFullTextQuery {
         query: String,
         columns: Vec<String>,
         boosts: Option<Vec<f64>>,
+        operator: String,
     ) -> napi::Result<Self> {
         let q = match boosts {
-            Some(boosts) => MultiMatchQuery::try_new_with_boosts(
+            Some(boosts) => MultiMatchQuery::try_new(query, columns)
-                query,
+                .and_then(|q| q.try_with_boosts(boosts.into_iter().map(|v| v as f32).collect())),
-                columns,
-                boosts.into_iter().map(|v| v as f32).collect(),
-            ),
             None => MultiMatchQuery::try_new(query, columns),
         }
         .map_err(|e| {
             napi::Error::from_reason(format!("Failed to create multi match query: {}", e))
         })?;
 
-        Ok(Self { inner: q.into() })
+        let operator = Operator::try_from(operator.as_str()).map_err(|e| {
+            napi::Error::from_reason(format!("Invalid operator for multi match query: {}", e))
+        })?;
+
+        Ok(Self {
+            inner: q.with_operator(operator).into(),
+        })
+    }
+
+    #[napi(factory)]
+    pub fn boolean_query(queries: Vec<(String, &JsFullTextQuery)>) -> napi::Result<Self> {
+        let mut sub_queries = Vec::with_capacity(queries.len());
+        for (occur, q) in queries {
+            let occur = Occur::try_from(occur.as_str())
+                .map_err(|e| napi::Error::from_reason(e.to_string()))?;
+            sub_queries.push((occur, q.inner.clone()));
+        }
+        Ok(Self {
+            inner: BooleanQuery::new(sub_queries).into(),
+        })
+    }
+
+    #[napi(getter)]
+    pub fn query_type(&self) -> String {
+        match self.inner {
+            FtsQuery::Match(_) => "match".to_string(),
+            FtsQuery::Phrase(_) => "phrase".to_string(),
+            FtsQuery::Boost(_) => "boost".to_string(),
+            FtsQuery::MultiMatch(_) => "multi_match".to_string(),
+            FtsQuery::Boolean(_) => "boolean".to_string(),
+        }
     }
 }
 

nodejs/src/table.rs

@@ -75,7 +75,7 @@ impl Table {
     }
 
     #[napi(catch_unwind)]
-    pub async fn add(&self, buf: Buffer, mode: String) -> napi::Result<()> {
+    pub async fn add(&self, buf: Buffer, mode: String) -> napi::Result<AddResult> {
         let batches = ipc_file_to_batches(buf.to_vec())
             .map_err(|e| napi::Error::from_reason(format!("Failed to read IPC file: {}", e)))?;
         let mut op = self.inner_ref()?.add(batches);
@@ -88,7 +88,8 @@ impl Table {
             return Err(napi::Error::from_reason(format!("Invalid mode: {}", mode)));
         };
 
-        op.execute().await.default_error()
+        let res = op.execute().await.default_error()?;
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
@@ -101,8 +102,9 @@ impl Table {
     }
 
     #[napi(catch_unwind)]
-    pub async fn delete(&self, predicate: String) -> napi::Result<()> {
+    pub async fn delete(&self, predicate: String) -> napi::Result<DeleteResult> {
-        self.inner_ref()?.delete(&predicate).await.default_error()
+        let res = self.inner_ref()?.delete(&predicate).await.default_error()?;
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
@@ -111,6 +113,7 @@ impl Table {
         index: Option<&Index>,
         column: String,
         replace: Option<bool>,
+        wait_timeout_s: Option<i64>,
     ) -> napi::Result<()> {
         let lancedb_index = if let Some(index) = index {
             index.consume()?
@@ -121,6 +124,10 @@ impl Table {
         if let Some(replace) = replace {
             builder = builder.replace(replace);
         }
+        if let Some(timeout) = wait_timeout_s {
+            builder =
+                builder.wait_timeout(std::time::Duration::from_secs(timeout.try_into().unwrap()));
+        }
         builder.execute().await.default_error()
     }
 
@@ -132,12 +139,38 @@ impl Table {
             .default_error()
     }
 
+    #[napi(catch_unwind)]
+    pub async fn prewarm_index(&self, index_name: String) -> napi::Result<()> {
+        self.inner_ref()?
+            .prewarm_index(&index_name)
+            .await
+            .default_error()
+    }
+
+    #[napi(catch_unwind)]
+    pub async fn wait_for_index(&self, index_names: Vec<String>, timeout_s: i64) -> Result<()> {
+        let timeout = std::time::Duration::from_secs(timeout_s.try_into().unwrap());
+        let index_names: Vec<&str> = index_names.iter().map(|s| s.as_str()).collect();
+        let slice: &[&str] = &index_names;
+
+        self.inner_ref()?
+            .wait_for_index(slice, timeout)
+            .await
+            .default_error()
+    }
+
+    #[napi(catch_unwind)]
+    pub async fn stats(&self) -> Result<TableStatistics> {
+        let stats = self.inner_ref()?.stats().await.default_error()?;
+        Ok(stats.into())
+    }
+
     #[napi(catch_unwind)]
     pub async fn update(
         &self,
         only_if: Option<String>,
         columns: Vec<(String, String)>,
-    ) -> napi::Result<u64> {
+    ) -> napi::Result<UpdateResult> {
         let mut op = self.inner_ref()?.update();
         if let Some(only_if) = only_if {
             op = op.only_if(only_if);
@@ -145,7 +178,8 @@ impl Table {
         for (column_name, value) in columns {
             op = op.column(column_name, value);
         }
-        op.execute().await.default_error()
+        let res = op.execute().await.default_error()?;
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
@@ -159,21 +193,28 @@ impl Table {
     }
 
     #[napi(catch_unwind)]
-    pub async fn add_columns(&self, transforms: Vec<AddColumnsSql>) -> napi::Result<()> {
+    pub async fn add_columns(
+        &self,
+        transforms: Vec<AddColumnsSql>,
+    ) -> napi::Result<AddColumnsResult> {
         let transforms = transforms
             .into_iter()
             .map(|sql| (sql.name, sql.value_sql))
             .collect::<Vec<_>>();
         let transforms = NewColumnTransform::SqlExpressions(transforms);
-        self.inner_ref()?
+        let res = self
+            .inner_ref()?
             .add_columns(transforms, None)
             .await
             .default_error()?;
-        Ok(())
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
-    pub async fn alter_columns(&self, alterations: Vec<ColumnAlteration>) -> napi::Result<()> {
+    pub async fn alter_columns(
+        &self,
+        alterations: Vec<ColumnAlteration>,
+    ) -> napi::Result<AlterColumnsResult> {
         for alteration in &alterations {
             if alteration.rename.is_none()
                 && alteration.nullable.is_none()
@@ -190,21 +231,23 @@ impl Table {
             .collect::<std::result::Result<Vec<_>, String>>()
             .map_err(napi::Error::from_reason)?;
 
-        self.inner_ref()?
+        let res = self
+            .inner_ref()?
             .alter_columns(&alterations)
             .await
             .default_error()?;
-        Ok(())
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
-    pub async fn drop_columns(&self, columns: Vec<String>) -> napi::Result<()> {
+    pub async fn drop_columns(&self, columns: Vec<String>) -> napi::Result<DropColumnsResult> {
         let col_refs = columns.iter().map(String::as_str).collect::<Vec<_>>();
-        self.inner_ref()?
+        let res = self
+            .inner_ref()?
             .drop_columns(&col_refs)
             .await
             .default_error()?;
-        Ok(())
+        Ok(res.into())
     }
 
     #[napi(catch_unwind)]
@@ -224,6 +267,14 @@ impl Table {
             .default_error()
     }
 
+    #[napi(catch_unwind)]
+    pub async fn checkout_tag(&self, tag: String) -> napi::Result<()> {
+        self.inner_ref()?
+            .checkout_tag(tag.as_str())
+            .await
+            .default_error()
+    }
+
     #[napi(catch_unwind)]
     pub async fn checkout_latest(&self) -> napi::Result<()> {
         self.inner_ref()?.checkout_latest().await.default_error()
@@ -256,6 +307,13 @@ impl Table {
         self.inner_ref()?.restore().await.default_error()
     }
 
+    #[napi(catch_unwind)]
+    pub async fn tags(&self) -> napi::Result<Tags> {
+        Ok(Tags {
+            inner: self.inner_ref()?.clone(),
+        })
+    }
+
     #[napi(catch_unwind)]
     pub async fn optimize(
         &self,
@@ -515,9 +573,257 @@ impl From<lancedb::index::IndexStatistics> for IndexStatistics {
     }
 }
 
+#[napi(object)]
+pub struct TableStatistics {
+    /// The total number of bytes in the table
+    pub total_bytes: i64,
+
+    /// The number of rows in the table
+    pub num_rows: i64,
+
+    /// The number of indices in the table
+    pub num_indices: i64,
+
+    /// Statistics on table fragments
+    pub fragment_stats: FragmentStatistics,
+}
+
+#[napi(object)]
+pub struct FragmentStatistics {
+    /// The number of fragments in the table
+    pub num_fragments: i64,
+
+    /// The number of uncompacted fragments in the table
+    pub num_small_fragments: i64,
+
+    /// Statistics on the number of rows in the table fragments
+    pub lengths: FragmentSummaryStats,
+}
+
+#[napi(object)]
+pub struct FragmentSummaryStats {
+    /// The number of rows in the fragment with the fewest rows
+    pub min: i64,
+
+    /// The number of rows in the fragment with the most rows
+    pub max: i64,
+
+    /// The mean number of rows in the fragments
+    pub mean: i64,
+
+    /// The 25th percentile of number of rows in the fragments
+    pub p25: i64,
+
+    /// The 50th percentile of number of rows in the fragments
+    pub p50: i64,
+
+    /// The 75th percentile of number of rows in the fragments
+    pub p75: i64,
+
+    /// The 99th percentile of number of rows in the fragments
+    pub p99: i64,
+}
+
+impl From<lancedb::table::TableStatistics> for TableStatistics {
+    fn from(v: lancedb::table::TableStatistics) -> Self {
+        Self {
+            total_bytes: v.total_bytes as i64,
+            num_rows: v.num_rows as i64,
+            num_indices: v.num_indices as i64,
+            fragment_stats: FragmentStatistics {
+                num_fragments: v.fragment_stats.num_fragments as i64,
+                num_small_fragments: v.fragment_stats.num_small_fragments as i64,
+                lengths: FragmentSummaryStats {
+                    min: v.fragment_stats.lengths.min as i64,
+                    max: v.fragment_stats.lengths.max as i64,
+                    mean: v.fragment_stats.lengths.mean as i64,
+                    p25: v.fragment_stats.lengths.p25 as i64,
+                    p50: v.fragment_stats.lengths.p50 as i64,
+                    p75: v.fragment_stats.lengths.p75 as i64,
+                    p99: v.fragment_stats.lengths.p99 as i64,
+                },
+            },
+        }
+    }
+}
+
 #[napi(object)]
 pub struct Version {
     pub version: i64,
     pub timestamp: i64,
     pub metadata: HashMap<String, String>,
 }
+
+#[napi(object)]
+pub struct UpdateResult {
+    pub rows_updated: i64,
+    pub version: i64,
+}
+
+impl From<lancedb::table::UpdateResult> for UpdateResult {
+    fn from(value: lancedb::table::UpdateResult) -> Self {
+        Self {
+            rows_updated: value.rows_updated as i64,
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct AddResult {
+    pub version: i64,
+}
+
+impl From<lancedb::table::AddResult> for AddResult {
+    fn from(value: lancedb::table::AddResult) -> Self {
+        Self {
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct DeleteResult {
+    pub version: i64,
+}
+
+impl From<lancedb::table::DeleteResult> for DeleteResult {
+    fn from(value: lancedb::table::DeleteResult) -> Self {
+        Self {
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct MergeResult {
+    pub version: i64,
+    pub num_inserted_rows: i64,
+    pub num_updated_rows: i64,
+    pub num_deleted_rows: i64,
+}
+
+impl From<lancedb::table::MergeResult> for MergeResult {
+    fn from(value: lancedb::table::MergeResult) -> Self {
+        Self {
+            version: value.version as i64,
+            num_inserted_rows: value.num_inserted_rows as i64,
+            num_updated_rows: value.num_updated_rows as i64,
+            num_deleted_rows: value.num_deleted_rows as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct AddColumnsResult {
+    pub version: i64,
+}
+
+impl From<lancedb::table::AddColumnsResult> for AddColumnsResult {
+    fn from(value: lancedb::table::AddColumnsResult) -> Self {
+        Self {
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct AlterColumnsResult {
+    pub version: i64,
+}
+
+impl From<lancedb::table::AlterColumnsResult> for AlterColumnsResult {
+    fn from(value: lancedb::table::AlterColumnsResult) -> Self {
+        Self {
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi(object)]
+pub struct DropColumnsResult {
+    pub version: i64,
+}
+
+impl From<lancedb::table::DropColumnsResult> for DropColumnsResult {
+    fn from(value: lancedb::table::DropColumnsResult) -> Self {
+        Self {
+            version: value.version as i64,
+        }
+    }
+}
+
+#[napi]
+pub struct TagContents {
+    pub version: i64,
+    pub manifest_size: i64,
+}
+
+#[napi]
+pub struct Tags {
+    inner: LanceDbTable,
+}
+
+#[napi]
+impl Tags {
+    #[napi]
+    pub async fn list(&self) -> napi::Result<HashMap<String, TagContents>> {
+        let rust_tags = self.inner.tags().await.default_error()?;
+        let tag_list = rust_tags.as_ref().list().await.default_error()?;
+        let tag_contents = tag_list
+            .into_iter()
+            .map(|(k, v)| {
+                (
+                    k,
+                    TagContents {
+                        version: v.version as i64,
+                        manifest_size: v.manifest_size as i64,
+                    },
+                )
+            })
+            .collect();
+
+        Ok(tag_contents)
+    }
+
+    #[napi]
+    pub async fn get_version(&self, tag: String) -> napi::Result<i64> {
+        let rust_tags = self.inner.tags().await.default_error()?;
+        rust_tags
+            .as_ref()
+            .get_version(tag.as_str())
+            .await
+            .map(|v| v as i64)
+            .default_error()
+    }
+
+    #[napi]
+    pub async unsafe fn create(&mut self, tag: String, version: i64) -> napi::Result<()> {
+        let mut rust_tags = self.inner.tags().await.default_error()?;
+        rust_tags
+            .as_mut()
+            .create(tag.as_str(), version as u64)
+            .await
+            .default_error()
+    }
+
+    #[napi]
+    pub async unsafe fn delete(&mut self, tag: String) -> napi::Result<()> {
+        let mut rust_tags = self.inner.tags().await.default_error()?;
+        rust_tags
+            .as_mut()
+            .delete(tag.as_str())
+            .await
+            .default_error()
+    }
+
+    #[napi]
+    pub async unsafe fn update(&mut self, tag: String, version: i64) -> napi::Result<()> {
+        let mut rust_tags = self.inner.tags().await.default_error()?;
+        rust_tags
+            .as_mut()
+            .update(tag.as_str(), version as u64)
+            .await
+            .default_error()
+    }
+}

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.22.0-beta.7"
+current_version = "0.24.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.22.0-beta.7"
+version = "0.24.0"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true
@@ -14,11 +14,11 @@ name = "_lancedb"
 crate-type = ["cdylib"]
 
 [dependencies]
-arrow = { version = "54.1", features = ["pyarrow"] }
+arrow = { version = "55.1", features = ["pyarrow"] }
 lancedb = { path = "../rust/lancedb", default-features = false }
 env_logger.workspace = true
-pyo3 = { version = "0.23", features = ["extension-module", "abi3-py39"] }
+pyo3 = { version = "0.24", features = ["extension-module", "abi3-py39"] }
-pyo3-async-runtimes = { version = "0.23", features = [
+pyo3-async-runtimes = { version = "0.24", features = [
     "attributes",
     "tokio-runtime",
 ] }
@@ -27,7 +27,7 @@ futures.workspace = true
 tokio = { version = "1.40", features = ["sync"] }
 
 [build-dependencies]
-pyo3-build-config = { version = "0.23", features = [
+pyo3-build-config = { version = "0.24", features = [
     "extension-module",
     "abi3-py39",
 ] }

python/pyproject.toml

@@ -7,7 +7,7 @@ dependencies = [
     "numpy",
     "overrides>=0.7",
     "packaging",
-    "pyarrow>=14",
+    "pyarrow>=16",
     "pydantic>=1.10",
     "tqdm>=4.27.0",
 ]
@@ -44,7 +44,7 @@ repository = "https://github.com/lancedb/lancedb"
 
 [project.optional-dependencies]
 pylance = [
-    "pylance>=0.23.2",
+    "pylance>=0.25",
 ]
 tests = [
     "aiohttp",
@@ -58,8 +58,9 @@ tests = [
     "polars>=0.19, <=1.3.0",
     "tantivy",
     "pyarrow-stubs",
-    "pylance>=0.23.2",
+    "pylance>=0.25",
     "requests",
+    "datafusion",
 ]
 dev = [
     "ruff",
@@ -77,6 +78,7 @@ embeddings = [
     "pillow",
     "open-clip-torch",
     "cohere",
+    "colpali-engine>=0.3.10",
     "huggingface_hub",
     "InstructorEmbedding",
     "google.generativeai",

python/python/lancedb/_lancedb.pyi

@@ -1,5 +1,5 @@
 from datetime import timedelta
-from typing import Dict, List, Optional, Tuple, Any, Union, Literal
+from typing import Dict, List, Optional, Tuple, Any, TypedDict, Union, Literal
 
 import pyarrow as pa
 
@@ -36,8 +36,10 @@ class Table:
     async def schema(self) -> pa.Schema: ...
     async def add(
         self, data: pa.RecordBatchReader, mode: Literal["append", "overwrite"]
-    ) -> None: ...
+    ) -> AddResult: ...
-    async def update(self, updates: Dict[str, str], where: Optional[str]) -> None: ...
+    async def update(
+        self, updates: Dict[str, str], where: Optional[str]
+    ) -> UpdateResult: ...
     async def count_rows(self, filter: Optional[str]) -> int: ...
     async def create_index(
         self,
@@ -47,23 +49,34 @@ class Table:
     ): ...
     async def list_versions(self) -> List[Dict[str, Any]]: ...
     async def version(self) -> int: ...
-    async def checkout(self, version: int): ...
+    async def checkout(self, version: Union[int, str]): ...
     async def checkout_latest(self): ...
-    async def restore(self, version: Optional[int] = None): ...
+    async def restore(self, version: Optional[Union[int, str]] = None): ...
     async def list_indices(self) -> list[IndexConfig]: ...
-    async def delete(self, filter: str): ...
+    async def delete(self, filter: str) -> DeleteResult: ...
-    async def add_columns(self, columns: list[tuple[str, str]]) -> None: ...
+    async def add_columns(self, columns: list[tuple[str, str]]) -> AddColumnsResult: ...
-    async def add_columns_with_schema(self, schema: pa.Schema) -> None: ...
+    async def add_columns_with_schema(self, schema: pa.Schema) -> AddColumnsResult: ...
-    async def alter_columns(self, columns: list[dict[str, Any]]) -> None: ...
+    async def alter_columns(
+        self, columns: list[dict[str, Any]]
+    ) -> AlterColumnsResult: ...
     async def optimize(
         self,
         *,
         cleanup_since_ms: Optional[int] = None,
         delete_unverified: Optional[bool] = None,
     ) -> OptimizeStats: ...
+    @property
+    def tags(self) -> Tags: ...
     def query(self) -> Query: ...
     def vector_search(self) -> VectorQuery: ...
 
+class Tags:
+    async def list(self) -> Dict[str, Tag]: ...
+    async def get_version(self, tag: str) -> int: ...
+    async def create(self, tag: str, version: int): ...
+    async def delete(self, tag: str): ...
+    async def update(self, tag: str, version: int): ...
+
 class IndexConfig:
     index_type: str
     columns: List[str]
@@ -130,6 +143,8 @@ class VectorQuery:
     def postfilter(self): ...
     def refine_factor(self, refine_factor: int): ...
     def nprobes(self, nprobes: int): ...
+    def minimum_nprobes(self, minimum_nprobes: int): ...
+    def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
     def nearest_to_text(self, query: dict) -> HybridQuery: ...
     def to_query_request(self) -> PyQueryRequest: ...
@@ -145,6 +160,8 @@ class HybridQuery:
     def distance_type(self, distance_type: str): ...
     def refine_factor(self, refine_factor: int): ...
     def nprobes(self, nprobes: int): ...
+    def minimum_nprobes(self, minimum_nprobes: int): ...
+    def maximum_nprobes(self, maximum_nprobes: int): ...
     def bypass_vector_index(self): ...
     def to_vector_query(self) -> VectorQuery: ...
     def to_fts_query(self) -> FTSQuery: ...
@@ -152,23 +169,21 @@ class HybridQuery:
     def get_with_row_id(self) -> bool: ...
     def to_query_request(self) -> PyQueryRequest: ...
 
-class PyFullTextSearchQuery:
+class FullTextQuery:
-    columns: Optional[List[str]]
+    pass
-    query: str
-    limit: Optional[int]
-    wand_factor: Optional[float]
 
 class PyQueryRequest:
     limit: Optional[int]
     offset: Optional[int]
     filter: Optional[Union[str, bytes]]
-    full_text_search: Optional[PyFullTextSearchQuery]
+    full_text_search: Optional[FullTextQuery]
     select: Optional[Union[str, List[str]]]
     fast_search: Optional[bool]
     with_row_id: Optional[bool]
     column: Optional[str]
     query_vector: Optional[List[pa.Array]]
-    nprobes: Optional[int]
+    minimum_nprobes: Optional[int]
+    maximum_nprobes: Optional[int]
     lower_bound: Optional[float]
     upper_bound: Optional[float]
     ef: Optional[int]
@@ -195,3 +210,32 @@ class RemovalStats:
 class OptimizeStats:
     compaction: CompactionStats
     prune: RemovalStats
+
+class Tag(TypedDict):
+    version: int
+    manifest_size: int
+
+class AddResult:
+    version: int
+
+class DeleteResult:
+    version: int
+
+class UpdateResult:
+    rows_updated: int
+    version: int
+
+class MergeResult:
+    version: int
+    num_updated_rows: int
+    num_inserted_rows: int
+    num_deleted_rows: int
+
+class AddColumnsResult:
+    version: int
+
+class AlterColumnsResult:
+    version: int
+
+class DropColumnsResult:
+    version: int

python/python/lancedb/common.py

@@ -9,7 +9,7 @@ import numpy as np
 import pyarrow as pa
 import pyarrow.dataset
 
-from .dependencies import pandas as pd
+from .dependencies import _check_for_pandas, pandas as pd
 
 DATA = Union[List[dict], "pd.DataFrame", pa.Table, Iterable[pa.RecordBatch]]
 VEC = Union[list, np.ndarray, pa.Array, pa.ChunkedArray]
@@ -63,7 +63,7 @@ def data_to_reader(
     data: DATA, schema: Optional[pa.Schema] = None
 ) -> pa.RecordBatchReader:
     """Convert various types of input into a RecordBatchReader"""
-    if pd is not None and isinstance(data, pd.DataFrame):
+    if _check_for_pandas(data) and isinstance(data, pd.DataFrame):
         return pa.Table.from_pandas(data, schema=schema).to_reader()
     elif isinstance(data, pa.Table):
         return data.to_reader()

python/python/lancedb/embeddings/__init__.py

@@ -19,3 +19,4 @@ from .imagebind import ImageBindEmbeddings
 from .jinaai import JinaEmbeddings
 from .watsonx import WatsonxEmbeddings
 from .voyageai import VoyageAIEmbeddingFunction
+from .colpali import ColPaliEmbeddings

python/python/lancedb/embeddings/colpali.py

@@ -0,0 +1,255 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+
+from functools import lru_cache
+from typing import List, Union, Optional, Any
+import numpy as np
+import io
+
+from ..util import attempt_import_or_raise
+from .base import EmbeddingFunction
+from .registry import register
+from .utils import TEXT, IMAGES, is_flash_attn_2_available
+
+
+@register("colpali")
+class ColPaliEmbeddings(EmbeddingFunction):
+    """
+    An embedding function that uses the ColPali engine for
+    multimodal multi-vector embeddings.
+
+    This embedding function supports ColQwen2.5 models, producing multivector outputs
+    for both text and image inputs. The output embeddings are lists of vectors, each
+    vector being 128-dimensional by default, represented as List[List[float]].
+
+    Parameters
+    ----------
+    model_name : str
+        The name of the model to use (e.g., "Metric-AI/ColQwen2.5-3b-multilingual-v1.0")
+    device : str
+        The device for inference (default "cuda:0").
+    dtype : str
+        Data type for model weights (default "bfloat16").
+    use_token_pooling : bool
+        Whether to use token pooling to reduce embedding size (default True).
+    pool_factor : int
+        Factor to reduce sequence length if token pooling is enabled (default 2).
+    quantization_config : Optional[BitsAndBytesConfig]
+        Quantization configuration for the model. (default None, bitsandbytes needed)
+    batch_size : int
+        Batch size for processing inputs (default 2).
+    """
+
+    model_name: str = "Metric-AI/ColQwen2.5-3b-multilingual-v1.0"
+    device: str = "auto"
+    dtype: str = "bfloat16"
+    use_token_pooling: bool = True
+    pool_factor: int = 2
+    quantization_config: Optional[Any] = None
+    batch_size: int = 2
+
+    _model = None
+    _processor = None
+    _token_pooler = None
+    _vector_dim = None
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        (
+            self._model,
+            self._processor,
+            self._token_pooler,
+        ) = self._load_model(
+            self.model_name,
+            self.dtype,
+            self.device,
+            self.use_token_pooling,
+            self.quantization_config,
+        )
+
+    @staticmethod
+    @lru_cache(maxsize=1)
+    def _load_model(
+        model_name: str,
+        dtype: str,
+        device: str,
+        use_token_pooling: bool,
+        quantization_config: Optional[Any],
+    ):
+        """
+        Initialize and cache the ColPali model, processor, and token pooler.
+        """
+        torch = attempt_import_or_raise("torch", "torch")
+        transformers = attempt_import_or_raise("transformers", "transformers")
+        colpali_engine = attempt_import_or_raise("colpali_engine", "colpali_engine")
+        from colpali_engine.compression.token_pooling import HierarchicalTokenPooler
+
+        if quantization_config is not None:
+            if not isinstance(quantization_config, transformers.BitsAndBytesConfig):
+                raise ValueError("quantization_config must be a BitsAndBytesConfig")
+
+        if dtype == "bfloat16":
+            torch_dtype = torch.bfloat16
+        elif dtype == "float16":
+            torch_dtype = torch.float16
+        elif dtype == "float64":
+            torch_dtype = torch.float64
+        else:
+            torch_dtype = torch.float32
+
+        model = colpali_engine.models.ColQwen2_5.from_pretrained(
+            model_name,
+            torch_dtype=torch_dtype,
+            device_map=device,
+            quantization_config=quantization_config
+            if quantization_config is not None
+            else None,
+            attn_implementation="flash_attention_2"
+            if is_flash_attn_2_available()
+            else None,
+        ).eval()
+        processor = colpali_engine.models.ColQwen2_5_Processor.from_pretrained(
+            model_name
+        )
+        token_pooler = HierarchicalTokenPooler() if use_token_pooling else None
+        return model, processor, token_pooler
+
+    def ndims(self):
+        """
+        Return the dimension of a vector in the multivector output (e.g., 128).
+        """
+        torch = attempt_import_or_raise("torch", "torch")
+        if self._vector_dim is None:
+            dummy_query = "test"
+            batch_queries = self._processor.process_queries([dummy_query]).to(
+                self._model.device
+            )
+            with torch.no_grad():
+                query_embeddings = self._model(**batch_queries)
+
+            if self.use_token_pooling and self._token_pooler is not None:
+                query_embeddings = self._token_pooler.pool_embeddings(
+                    query_embeddings,
+                    pool_factor=self.pool_factor,
+                    padding=True,
+                    padding_side=self._processor.tokenizer.padding_side,
+                )
+
+            self._vector_dim = query_embeddings[0].shape[-1]
+        return self._vector_dim
+
+    def _process_embeddings(self, embeddings):
+        """
+        Format model embeddings into List[List[float]].
+        Use token pooling if enabled.
+        """
+        torch = attempt_import_or_raise("torch", "torch")
+        if self.use_token_pooling and self._token_pooler is not None:
+            embeddings = self._token_pooler.pool_embeddings(
+                embeddings,
+                pool_factor=self.pool_factor,
+                padding=True,
+                padding_side=self._processor.tokenizer.padding_side,
+            )
+
+        if isinstance(embeddings, torch.Tensor):
+            tensors = embeddings.detach().cpu()
+            if tensors.dtype == torch.bfloat16:
+                tensors = tensors.to(torch.float32)
+            return (
+                tensors.numpy()
+                .astype(np.float64 if self.dtype == "float64" else np.float32)
+                .tolist()
+            )
+        return []
+
+    def generate_text_embeddings(self, text: TEXT) -> List[List[List[float]]]:
+        """
+        Generate embeddings for text input.
+        """
+        torch = attempt_import_or_raise("torch", "torch")
+        text = self.sanitize_input(text)
+        all_embeddings = []
+
+        for i in range(0, len(text), self.batch_size):
+            batch_text = text[i : i + self.batch_size]
+            batch_queries = self._processor.process_queries(batch_text).to(
+                self._model.device
+            )
+            with torch.no_grad():
+                query_embeddings = self._model(**batch_queries)
+            all_embeddings.extend(self._process_embeddings(query_embeddings))
+        return all_embeddings
+
+    def _prepare_images(self, images: IMAGES) -> List:
+        """
+        Convert image inputs to PIL Images.
+        """
+        PIL = attempt_import_or_raise("PIL", "pillow")
+        requests = attempt_import_or_raise("requests", "requests")
+        images = self.sanitize_input(images)
+        pil_images = []
+        try:
+            for image in images:
+                if isinstance(image, str):
+                    if image.startswith(("http://", "https://")):
+                        response = requests.get(image, timeout=10)
+                        response.raise_for_status()
+                        pil_images.append(PIL.Image.open(io.BytesIO(response.content)))
+                    else:
+                        with PIL.Image.open(image) as im:
+                            pil_images.append(im.copy())
+                elif isinstance(image, bytes):
+                    pil_images.append(PIL.Image.open(io.BytesIO(image)))
+                else:
+                    # Assume it's a PIL Image; will raise if invalid
+                    pil_images.append(image)
+        except Exception as e:
+            raise ValueError(f"Failed to process image: {e}")
+
+        return pil_images
+
+    def generate_image_embeddings(self, images: IMAGES) -> List[List[List[float]]]:
+        """
+        Generate embeddings for a batch of images.
+        """
+        torch = attempt_import_or_raise("torch", "torch")
+        pil_images = self._prepare_images(images)
+        all_embeddings = []
+
+        for i in range(0, len(pil_images), self.batch_size):
+            batch_images = pil_images[i : i + self.batch_size]
+            batch_images = self._processor.process_images(batch_images).to(
+                self._model.device
+            )
+            with torch.no_grad():
+                image_embeddings = self._model(**batch_images)
+            all_embeddings.extend(self._process_embeddings(image_embeddings))
+        return all_embeddings
+
+    def compute_query_embeddings(
+        self, query: Union[str, IMAGES], *args, **kwargs
+    ) -> List[List[List[float]]]:
+        """
+        Compute embeddings for a single user query (text only).
+        """
+        if not isinstance(query, str):
+            raise ValueError(
+                "Query must be a string, image to image search is not supported"
+            )
+        return self.generate_text_embeddings([query])
+
+    def compute_source_embeddings(
+        self, images: IMAGES, *args, **kwargs
+    ) -> List[List[List[float]]]:
+        """
+        Compute embeddings for a batch of source images.
+
+        Parameters
+        ----------
+        images : Union[str, bytes, List, pa.Array, pa.ChunkedArray, np.ndarray]
+            Batch of images (paths, URLs, bytes, or PIL Images).
+        """
+        images = self.sanitize_input(images)
+        return self.generate_image_embeddings(images)

python/python/lancedb/embeddings/utils.py

@@ -18,6 +18,7 @@ import numpy as np
 import pyarrow as pa
 
 from ..dependencies import pandas as pd
+from ..util import attempt_import_or_raise
 
 
 # ruff: noqa: PERF203
@@ -275,3 +276,12 @@ def url_retrieve(url: str):
 def api_key_not_found_help(provider):
     logging.error("Could not find API key for %s", provider)
     raise ValueError(f"Please set the {provider.upper()}_API_KEY environment variable.")
+
+
+def is_flash_attn_2_available():
+    try:
+        attempt_import_or_raise("flash_attn", "flash_attn")
+
+        return True
+    except ImportError:
+        return False

python/python/lancedb/index.py

@@ -102,7 +102,7 @@ class FTS:
 
     Attributes
     ----------
-    with_position : bool, default True
+    with_position : bool, default False
         Whether to store the position of the token in the document. Setting this
         to False can reduce the size of the index and improve indexing speed,
         but it will disable support for phrase queries.
@@ -118,25 +118,25 @@ class FTS:
         ignored.
     lower_case : bool, default True
         Whether to convert the token to lower case. This makes queries case-insensitive.
-    stem : bool, default False
+    stem : bool, default True
         Whether to stem the token. Stemming reduces words to their root form.
         For example, in English "running" and "runs" would both be reduced to "run".
-    remove_stop_words : bool, default False
+    remove_stop_words : bool, default True
         Whether to remove stop words. Stop words are common words that are often
         removed from text before indexing. For example, in English "the" and "and".
-    ascii_folding : bool, default False
+    ascii_folding : bool, default True
         Whether to fold ASCII characters. This converts accented characters to
         their ASCII equivalent. For example, "café" would be converted to "cafe".
     """
 
-    with_position: bool = True
+    with_position: bool = False
     base_tokenizer: Literal["simple", "raw", "whitespace"] = "simple"
     language: str = "English"
     max_token_length: Optional[int] = 40
     lower_case: bool = True
-    stem: bool = False
+    stem: bool = True
-    remove_stop_words: bool = False
+    remove_stop_words: bool = True
-    ascii_folding: bool = False
+    ascii_folding: bool = True
 
 
 @dataclass

python/python/lancedb/merge.py

@@ -4,10 +4,14 @@
 
 from __future__ import annotations
 
+from datetime import timedelta
 from typing import TYPE_CHECKING, List, Optional
 
 if TYPE_CHECKING:
     from .common import DATA
+    from ._lancedb import (
+        MergeInsertResult,
+    )
 
 
 class LanceMergeInsertBuilder(object):
@@ -28,6 +32,7 @@ class LanceMergeInsertBuilder(object):
         self._when_not_matched_insert_all = False
         self._when_not_matched_by_source_delete = False
         self._when_not_matched_by_source_condition = None
+        self._timeout = None
 
     def when_matched_update_all(
         self, *, where: Optional[str] = None
@@ -78,7 +83,8 @@ class LanceMergeInsertBuilder(object):
         new_data: DATA,
         on_bad_vectors: str = "error",
         fill_value: float = 0.0,
-    ):
+        timeout: Optional[timedelta] = None,
+    ) -> MergeInsertResult:
         """
         Executes the merge insert operation
 
@@ -95,5 +101,24 @@ class LanceMergeInsertBuilder(object):
             One of "error", "drop", "fill".
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        timeout: Optional[timedelta], default None
+            Maximum time to run the operation before cancelling it.
+
+            By default, there is a 30-second timeout that is only enforced after the
+            first attempt. This is to prevent spending too long retrying to resolve
+            conflicts. For example, if a write attempt takes 20 seconds and fails,
+            the second attempt will be cancelled after 10 seconds, hitting the
+            30-second timeout. However, a write that takes one hour and succeeds on the
+            first attempt will not be cancelled.
+
+            When this is set, the timeout is enforced on all attempts, including
+            the first.
+
+        Returns
+        -------
+        MergeInsertResult
+            version: the new version number of the table after doing merge insert.
         """
+        if timeout is not None:
+            self._timeout = timeout
         return self._table._do_merge(self, new_data, on_bad_vectors, fill_value)

python/python/lancedb/pydantic.py

@@ -152,6 +152,104 @@ def Vector(
     return FixedSizeList
 
 
+def MultiVector(
+    dim: int, value_type: pa.DataType = pa.float32(), nullable: bool = True
+) -> Type:
+    """Pydantic MultiVector Type for multi-vector embeddings.
+
+    This type represents a list of vectors, each with the same dimension.
+    Useful for models that produce multiple embeddings per input, like ColPali.
+
+    Parameters
+    ----------
+    dim : int
+        The dimension of each vector in the multi-vector.
+    value_type : pyarrow.DataType, optional
+        The value type of the vectors, by default pa.float32()
+    nullable : bool, optional
+        Whether the multi-vector is nullable, by default it is True.
+
+    Examples
+    --------
+
+    >>> import pydantic
+    >>> from lancedb.pydantic import MultiVector
+    ...
+    >>> class MyModel(pydantic.BaseModel):
+    ...     id: int
+    ...     text: str
+    ...     embeddings: MultiVector(128)  # List of 128-dimensional vectors
+    >>> schema = pydantic_to_schema(MyModel)
+    >>> assert schema == pa.schema([
+    ...     pa.field("id", pa.int64(), False),
+    ...     pa.field("text", pa.utf8(), False),
+    ...     pa.field("embeddings", pa.list_(pa.list_(pa.float32(), 128)))
+    ... ])
+    """
+
+    class MultiVectorList(list, FixedSizeListMixin):
+        def __repr__(self):
+            return f"MultiVector(dim={dim})"
+
+        @staticmethod
+        def nullable() -> bool:
+            return nullable
+
+        @staticmethod
+        def dim() -> int:
+            return dim
+
+        @staticmethod
+        def value_arrow_type() -> pa.DataType:
+            return value_type
+
+        @staticmethod
+        def is_multi_vector() -> bool:
+            return True
+
+        @classmethod
+        def __get_pydantic_core_schema__(
+            cls, _source_type: Any, _handler: pydantic.GetCoreSchemaHandler
+        ) -> CoreSchema:
+            return core_schema.no_info_after_validator_function(
+                cls,
+                core_schema.list_schema(
+                    items_schema=core_schema.list_schema(
+                        min_length=dim,
+                        max_length=dim,
+                        items_schema=core_schema.float_schema(),
+                    ),
+                ),
+            )
+
+        @classmethod
+        def __get_validators__(cls) -> Generator[Callable, None, None]:
+            yield cls.validate
+
+        # For pydantic v1
+        @classmethod
+        def validate(cls, v):
+            if not isinstance(v, (list, range)):
+                raise TypeError("A list of vectors is needed")
+            for vec in v:
+                if not isinstance(vec, (list, range, np.ndarray)) or len(vec) != dim:
+                    raise TypeError(f"Each vector must be a list of {dim} numbers")
+            return cls(v)
+
+        if PYDANTIC_VERSION.major < 2:
+
+            @classmethod
+            def __modify_schema__(cls, field_schema: Dict[str, Any]):
+                field_schema["items"] = {
+                    "type": "array",
+                    "items": {"type": "number"},
+                    "minItems": dim,
+                    "maxItems": dim,
+                }
+
+    return MultiVectorList
+
+
 def _py_type_to_arrow_type(py_type: Type[Any], field: FieldInfo) -> pa.DataType:
     """Convert a field with native Python type to Arrow data type.
 
@@ -206,6 +304,9 @@ def _pydantic_type_to_arrow_type(tp: Any, field: FieldInfo) -> pa.DataType:
             fields = _pydantic_model_to_fields(tp)
             return pa.struct(fields)
         if issubclass(tp, FixedSizeListMixin):
+            if getattr(tp, "is_multi_vector", lambda: False)():
+                return pa.list_(pa.list_(tp.value_arrow_type(), tp.dim()))
+            # For regular Vector
             return pa.list_(tp.value_arrow_type(), tp.dim())
     return _py_type_to_arrow_type(tp, field)
 
@@ -314,6 +415,7 @@ class LanceModel(pydantic.BaseModel):
     >>> table.add([
     ...     TestModel(name="test", vector=[1.0, 2.0])
     ... ])
+    AddResult(version=2)
     >>> table.search([0., 0.]).limit(1).to_pydantic(TestModel)
     [TestModel(name='test', vector=FixedSizeList(dim=2))]
     """

python/python/lancedb/query.py

@@ -4,7 +4,6 @@
 from __future__ import annotations
 
 from abc import ABC, abstractmethod
-import abc
 from concurrent.futures import ThreadPoolExecutor
 from enum import Enum
 from datetime import timedelta
@@ -28,6 +27,8 @@ import pyarrow.compute as pc
 import pyarrow.fs as pa_fs
 import pydantic
 
+from lancedb.pydantic import PYDANTIC_VERSION
+
 from . import __version__
 from .arrow import AsyncRecordBatchReader
 from .dependencies import pandas as pd
@@ -86,15 +87,28 @@ def ensure_vector_query(
         return val
 
 
-class FullTextQueryType(Enum):
+class FullTextQueryType(str, Enum):
     MATCH = "match"
     MATCH_PHRASE = "match_phrase"
     BOOST = "boost"
     MULTI_MATCH = "multi_match"
+    BOOLEAN = "boolean"
 
 
-class FullTextQuery(abc.ABC, pydantic.BaseModel):
+class FullTextOperator(str, Enum):
-    @abc.abstractmethod
+    AND = "AND"
+    OR = "OR"
+
+
+class Occur(str, Enum):
+    SHOULD = "SHOULD"
+    MUST = "MUST"
+    MUST_NOT = "MUST_NOT"
+
+
+@pydantic.dataclasses.dataclass
+class FullTextQuery(ABC):
+    @abstractmethod
     def query_type(self) -> FullTextQueryType:
         """
         Get the query type of the query.
@@ -104,193 +118,178 @@ class FullTextQuery(abc.ABC, pydantic.BaseModel):
         str
             The type of the query.
         """
+        pass
 
-    @abc.abstractmethod
+    def __and__(self, other: "FullTextQuery") -> "FullTextQuery":
-    def to_dict(self) -> dict:
         """
-        Convert the query to a dictionary.
+        Combine two queries with a logical AND operation.
-
-        Returns
-        -------
-        dict
-            The query as a dictionary.
-        """
-
-
-class MatchQuery(FullTextQuery):
-    query: str
-    column: str
-    boost: float = 1.0
-    fuzziness: int = 0
-    max_expansions: int = 50
-
-    def __init__(
-        self,
-        query: str,
-        column: str,
-        *,
-        boost: float = 1.0,
-        fuzziness: int = 0,
-        max_expansions: int = 50,
-    ):
-        """
-        Match query for full-text search.
 
         Parameters
         ----------
-        query : str
+        other : FullTextQuery
-            The query string to match against.
+            The other query to combine with.
-        column : str
+
-            The name of the column to match against.
+        Returns
-        boost : float, default 1.0
+        -------
-            The boost factor for the query.
+        FullTextQuery
-            The score of each matching document is multiplied by this value.
+            A new query that combines both queries with AND.
-        fuzziness : int, optional
-            The maximum edit distance for each term in the match query.
-            Defaults to 0 (exact match).
-            If None, fuzziness is applied automatically by the rules:
-                - 0 for terms with length <= 2
-                - 1 for terms with length <= 5
-                - 2 for terms with length > 5
-        max_expansions : int, optional
-            The maximum number of terms to consider for fuzzy matching.
-            Defaults to 50.
         """
-        super().__init__(
+        return BooleanQuery([(Occur.MUST, self), (Occur.MUST, other)])
-            query=query,
+
-            column=column,
+    def __or__(self, other: "FullTextQuery") -> "FullTextQuery":
-            boost=boost,
+        """
-            fuzziness=fuzziness,
+        Combine two queries with a logical OR operation.
-            max_expansions=max_expansions,
+
-        )
+        Parameters
+        ----------
+        other : FullTextQuery
+            The other query to combine with.
+
+        Returns
+        -------
+        FullTextQuery
+            A new query that combines both queries with OR.
+        """
+        return BooleanQuery([(Occur.SHOULD, self), (Occur.SHOULD, other)])
+
+
+@pydantic.dataclasses.dataclass
+class MatchQuery(FullTextQuery):
+    """
+    Match query for full-text search.
+
+    Parameters
+    ----------
+    query : str
+        The query string to match against.
+    column : str
+        The name of the column to match against.
+    boost : float, default 1.0
+        The boost factor for the query.
+        The score of each matching document is multiplied by this value.
+    fuzziness : int, optional
+        The maximum edit distance for each term in the match query.
+        Defaults to 0 (exact match).
+        If None, fuzziness is applied automatically by the rules:
+            - 0 for terms with length <= 2
+            - 1 for terms with length <= 5
+            - 2 for terms with length > 5
+    max_expansions : int, optional
+        The maximum number of terms to consider for fuzzy matching.
+        Defaults to 50.
+    operator : FullTextOperator, default OR
+        The operator to use for combining the query results.
+        Can be either `AND` or `OR`.
+        If `AND`, all terms in the query must match.
+        If `OR`, at least one term in the query must match.
+    prefix_length : int, optional
+        The number of beginning characters being unchanged for fuzzy matching.
+        This is useful to achieve prefix matching.
+    """
+
+    query: str
+    column: str
+    boost: float = pydantic.Field(1.0, kw_only=True)
+    fuzziness: int = pydantic.Field(0, kw_only=True)
+    max_expansions: int = pydantic.Field(50, kw_only=True)
+    operator: FullTextOperator = pydantic.Field(FullTextOperator.OR, kw_only=True)
+    prefix_length: int = pydantic.Field(0, kw_only=True)
 
     def query_type(self) -> FullTextQueryType:
         return FullTextQueryType.MATCH
 
-    def to_dict(self) -> dict:
-        return {
-            "match": {
-                self.column: {
-                    "query": self.query,
-                    "boost": self.boost,
-                    "fuzziness": self.fuzziness,
-                    "max_expansions": self.max_expansions,
-                }
-            }
-        }
-
 
+@pydantic.dataclasses.dataclass
 class PhraseQuery(FullTextQuery):
+    """
+    Phrase query for full-text search.
+
+    Parameters
+    ----------
+    query : str
+        The query string to match against.
+    column : str
+        The name of the column to match against.
+    """
+
     query: str
     column: str
-
+    slop: int = pydantic.Field(0, kw_only=True)
-    def __init__(self, query: str, column: str):
-        """
-        Phrase query for full-text search.
-
-        Parameters
-        ----------
-        query : str
-            The query string to match against.
-        column : str
-            The name of the column to match against.
-        """
-        super().__init__(query=query, column=column)
 
     def query_type(self) -> FullTextQueryType:
         return FullTextQueryType.MATCH_PHRASE
 
-    def to_dict(self) -> dict:
-        return {
-            "match_phrase": {
-                self.column: self.query,
-            }
-        }
-
 
+@pydantic.dataclasses.dataclass
 class BoostQuery(FullTextQuery):
+    """
+    Boost query for full-text search.
+
+    Parameters
+    ----------
+    positive : dict
+        The positive query object.
+    negative : dict
+        The negative query object.
+    negative_boost : float, default 0.5
+        The boost factor for the negative query.
+    """
+
     positive: FullTextQuery
     negative: FullTextQuery
-    negative_boost: float = 0.5
+    negative_boost: float = pydantic.Field(0.5, kw_only=True)
-
-    def __init__(
-        self,
-        positive: FullTextQuery,
-        negative: FullTextQuery,
-        *,
-        negative_boost: float = 0.5,
-    ):
-        """
-        Boost query for full-text search.
-
-        Parameters
-        ----------
-        positive : dict
-            The positive query object.
-        negative : dict
-            The negative query object.
-        negative_boost : float
-            The boost factor for the negative query.
-        """
-        super().__init__(
-            positive=positive, negative=negative, negative_boost=negative_boost
-        )
 
     def query_type(self) -> FullTextQueryType:
         return FullTextQueryType.BOOST
 
-    def to_dict(self) -> dict:
-        return {
-            "boost": {
-                "positive": self.positive.to_dict(),
-                "negative": self.negative.to_dict(),
-                "negative_boost": self.negative_boost,
-            }
-        }
-
 
+@pydantic.dataclasses.dataclass
 class MultiMatchQuery(FullTextQuery):
+    """
+    Multi-match query for full-text search.
+
+    Parameters
+    ----------
+    query : str | list[Query]
+        If a string, the query string to match against.
+    columns : list[str]
+        The list of columns to match against.
+    boosts : list[float], optional
+        The list of boost factors for each column. If not provided,
+        all columns will have the same boost factor.
+    operator : FullTextOperator, default OR
+        The operator to use for combining the query results.
+        Can be either `AND` or `OR`.
+        It would be applied to all columns individually.
+        For example, if the operator is `AND`,
+        then the query "hello world" is equal to
+        `match("hello AND world", column1) OR match("hello AND world", column2)`.
+    """
+
     query: str
     columns: list[str]
-    boosts: list[float]
+    boosts: Optional[list[float]] = pydantic.Field(None, kw_only=True)
-
+    operator: FullTextOperator = pydantic.Field(FullTextOperator.OR, kw_only=True)
-    def __init__(
-        self,
-        query: str,
-        columns: list[str],
-        *,
-        boosts: Optional[list[float]] = None,
-    ):
-        """
-        Multi-match query for full-text search.
-
-        Parameters
-        ----------
-        query : str
-            The query string to match against.
-
-        columns : list[str]
-            The list of columns to match against.
-
-        boosts : list[float], optional
-            The list of boost factors for each column. If not provided,
-            all columns will have the same boost factor.
-        """
-        if boosts is None:
-            boosts = [1.0] * len(columns)
-        super().__init__(query=query, columns=columns, boosts=boosts)
 
     def query_type(self) -> FullTextQueryType:
         return FullTextQueryType.MULTI_MATCH
 
-    def to_dict(self) -> dict:
+
-        return {
+@pydantic.dataclasses.dataclass
-            "multi_match": {
+class BooleanQuery(FullTextQuery):
-                "query": self.query,
+    """
-                "columns": self.columns,
+    Boolean query for full-text search.
-                "boost": self.boosts,
+
-            }
+    Parameters
-        }
+    ----------
+    queries : list[tuple(Occur, FullTextQuery)]
+        The list of queries with their occurrence requirements.
+    """
+
+    queries: list[tuple[Occur, FullTextQuery]]
+
+    def query_type(self) -> FullTextQueryType:
+        return FullTextQueryType.BOOLEAN
 
 
 class FullTextSearchQuery(pydantic.BaseModel):
@@ -443,8 +442,18 @@ class Query(pydantic.BaseModel):
     # which columns to return in the results
     columns: Optional[Union[List[str], Dict[str, str]]] = None
 
-    # number of IVF partitions to search
+    # minimum number of IVF partitions to search
-    nprobes: Optional[int] = None
+    #
+    # If None then a default value (20) will be used.
+    minimum_nprobes: Optional[int] = None
+
+    # maximum number of IVF partitions to search
+    #
+    # If None then a default value (20) will be used.
+    #
+    # If 0 then no limit will be applied and all partitions could be searched
+    # if needed to satisfy the limit.
+    maximum_nprobes: Optional[int] = None
 
     # lower bound for distance search
     lower_bound: Optional[float] = None
@@ -482,7 +491,8 @@ class Query(pydantic.BaseModel):
         query.vector_column = req.column
         query.vector = req.query_vector
         query.distance_type = req.distance_type
-        query.nprobes = req.nprobes
+        query.minimum_nprobes = req.minimum_nprobes
+        query.maximum_nprobes = req.maximum_nprobes
         query.lower_bound = req.lower_bound
         query.upper_bound = req.upper_bound
         query.ef = req.ef
@@ -491,17 +501,19 @@ class Query(pydantic.BaseModel):
         query.postfilter = req.postfilter
         if req.full_text_search is not None:
             query.full_text_query = FullTextSearchQuery(
-                columns=req.full_text_search.columns,
+                columns=None,
-                query=req.full_text_search.query,
+                query=req.full_text_search,
-                limit=req.full_text_search.limit,
-                wand_factor=req.full_text_search.wand_factor,
             )
         return query
 
-    class Config:
+    # This tells pydantic to allow custom types (needed for the `vector` query since
-        # This tells pydantic to allow custom types (needed for the `vector` query since
+    # pa.Array wouln't be allowed otherwise)
-        # pa.Array wouln't be allowed otherwise)
+    if PYDANTIC_VERSION.major < 2:  # Pydantic 1.x compat
-        arbitrary_types_allowed = True
+
+        class Config:
+            arbitrary_types_allowed = True
+    else:
+        model_config = {"arbitrary_types_allowed": True}
 
 
 class LanceQueryBuilder(ABC):
@@ -1041,7 +1053,8 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         super().__init__(table)
         self._query = query
         self._distance_type = None
-        self._nprobes = None
+        self._minimum_nprobes = None
+        self._maximum_nprobes = None
         self._lower_bound = None
         self._upper_bound = None
         self._refine_factor = None
@@ -1104,6 +1117,10 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         See discussion in [Querying an ANN Index][querying-an-ann-index] for
         tuning advice.
 
+        This method sets both the minimum and maximum number of probes to the same
+        value. See `minimum_nprobes` and `maximum_nprobes` for more fine-grained
+        control.
+
         Parameters
         ----------
         nprobes: int
@@ -1114,7 +1131,36 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
         LanceVectorQueryBuilder
             The LanceQueryBuilder object.
         """
-        self._nprobes = nprobes
+        self._minimum_nprobes = nprobes
+        self._maximum_nprobes = nprobes
+        return self
+
+    def minimum_nprobes(self, minimum_nprobes: int) -> LanceVectorQueryBuilder:
+        """Set the minimum number of probes to use.
+
+        See `nprobes` for more details.
+
+        These partitions will be searched on every vector query and will increase recall
+        at the expense of latency.
+        """
+        self._minimum_nprobes = minimum_nprobes
+        return self
+
+    def maximum_nprobes(self, maximum_nprobes: int) -> LanceVectorQueryBuilder:
+        """Set the maximum number of probes to use.
+
+        See `nprobes` for more details.
+
+        If this value is greater than `minimum_nprobes` then the excess partitions
+        will be searched only if we have not found enough results.
+
+        This can be useful when there is a narrow filter to allow these queries to
+        spend more time searching and avoid potential false negatives.
+
+        If this value is 0 then no limit will be applied and all partitions could be
+        searched if needed to satisfy the limit.
+        """
+        self._maximum_nprobes = maximum_nprobes
         return self
 
     def distance_range(
@@ -1218,7 +1264,8 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
             limit=self._limit,
             distance_type=self._distance_type,
             columns=self._columns,
-            nprobes=self._nprobes,
+            minimum_nprobes=self._minimum_nprobes,
+            maximum_nprobes=self._maximum_nprobes,
             lower_bound=self._lower_bound,
             upper_bound=self._upper_bound,
             refine_factor=self._refine_factor,
@@ -1404,10 +1451,13 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
 
         query = self._query
         if self._phrase_query:
-            raise NotImplementedError(
+            if isinstance(query, str):
-                "Phrase query is not yet supported in Lance FTS. "
+                if not query.startswith('"') or not query.endswith('"'):
-                "Use tantivy-based index instead for now."
+                    query = f'"{query}"'
-            )
+            elif isinstance(query, FullTextQuery) and not isinstance(
+                query, PhraseQuery
+            ):
+                raise TypeError("Please use PhraseQuery for phrase queries.")
         query = self.to_query_object()
         results = self._table._execute_query(query, timeout=timeout)
         results = results.read_all()
@@ -1582,10 +1632,13 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         self._fts_columns = fts_columns
         self._norm = None
         self._reranker = None
-        self._nprobes = None
+        self._minimum_nprobes = None
+        self._maximum_nprobes = None
         self._refine_factor = None
         self._distance_type = None
         self._phrase_query = None
+        self._lower_bound = None
+        self._upper_bound = None
 
     def _validate_query(self, query, vector=None, text=None):
         if query is not None and (vector is not None or text is not None):
@@ -1628,47 +1681,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         raise NotImplementedError("to_query_object not yet supported on a hybrid query")
 
     def to_arrow(self, *, timeout: Optional[timedelta] = None) -> pa.Table:
-        vector_query, fts_query = self._validate_query(
+        self._create_query_builders()
-            self._query, self._vector, self._text
-        )
-        self._fts_query = LanceFtsQueryBuilder(
-            self._table, fts_query, fts_columns=self._fts_columns
-        )
-        vector_query = self._query_to_vector(
-            self._table, vector_query, self._vector_column
-        )
-        self._vector_query = LanceVectorQueryBuilder(
-            self._table, vector_query, self._vector_column
-        )
-
-        if self._limit:
-            self._vector_query.limit(self._limit)
-            self._fts_query.limit(self._limit)
-        if self._columns:
-            self._vector_query.select(self._columns)
-            self._fts_query.select(self._columns)
-        if self._where:
-            self._vector_query.where(self._where, self._postfilter)
-            self._fts_query.where(self._where, self._postfilter)
-        if self._with_row_id:
-            self._vector_query.with_row_id(True)
-            self._fts_query.with_row_id(True)
-        if self._phrase_query:
-            self._fts_query.phrase_query(True)
-        if self._distance_type:
-            self._vector_query.metric(self._distance_type)
-        if self._nprobes:
-            self._vector_query.nprobes(self._nprobes)
-        if self._refine_factor:
-            self._vector_query.refine_factor(self._refine_factor)
-        if self._ef:
-            self._vector_query.ef(self._ef)
-        if self._bypass_vector_index:
-            self._vector_query.bypass_vector_index()
-
-        if self._reranker is None:
-            self._reranker = RRFReranker()
-
         with ThreadPoolExecutor() as executor:
             fts_future = executor.submit(
                 self._fts_query.with_row_id(True).to_arrow, timeout=timeout
@@ -1852,7 +1865,24 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         LanceHybridQueryBuilder
             The LanceHybridQueryBuilder object.
         """
-        self._nprobes = nprobes
+        self._minimum_nprobes = nprobes
+        self._maximum_nprobes = nprobes
+        return self
+
+    def minimum_nprobes(self, minimum_nprobes: int) -> LanceHybridQueryBuilder:
+        """Set the minimum number of probes to use.
+
+        See `nprobes` for more details.
+        """
+        self._minimum_nprobes = minimum_nprobes
+        return self
+
+    def maximum_nprobes(self, maximum_nprobes: int) -> LanceHybridQueryBuilder:
+        """Set the maximum number of probes to use.
+
+        See `nprobes` for more details.
+        """
+        self._maximum_nprobes = maximum_nprobes
         return self
 
     def distance_range(
@@ -1991,6 +2021,114 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
         self._bypass_vector_index = True
         return self
 
+    def explain_plan(self, verbose: Optional[bool] = False) -> str:
+        """Return the execution plan for this query.
+
+        Examples
+        --------
+        >>> import lancedb
+        >>> db = lancedb.connect("./.lancedb")
+        >>> table = db.create_table("my_table", [{"vector": [99.0, 99]}])
+        >>> query = [100, 100]
+        >>> plan = table.search(query).explain_plan(True)
+        >>> print(plan) # doctest: +ELLIPSIS, +NORMALIZE_WHITESPACE
+        ProjectionExec: expr=[vector@0 as vector, _distance@2 as _distance]
+        GlobalLimitExec: skip=0, fetch=10
+          FilterExec: _distance@2 IS NOT NULL
+            SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], preserve_partitioning=[false]
+              KNNVectorDistance: metric=l2
+                LanceScan: uri=..., projection=[vector], row_id=true, row_addr=false, ordered=false
+
+        Parameters
+        ----------
+        verbose : bool, default False
+            Use a verbose output format.
+
+        Returns
+        -------
+        plan : str
+        """  # noqa: E501
+        self._create_query_builders()
+
+        results = ["Vector Search Plan:"]
+        results.append(
+            self._table._explain_plan(
+                self._vector_query.to_query_object(), verbose=verbose
+            )
+        )
+        results.append("FTS Search Plan:")
+        results.append(
+            self._table._explain_plan(
+                self._fts_query.to_query_object(), verbose=verbose
+            )
+        )
+        return "\n".join(results)
+
+    def analyze_plan(self):
+        """Execute the query and display with runtime metrics.
+
+        Returns
+        -------
+        plan : str
+        """
+        self._create_query_builders()
+
+        results = ["Vector Search Plan:"]
+        results.append(self._table._analyze_plan(self._vector_query.to_query_object()))
+        results.append("FTS Search Plan:")
+        results.append(self._table._analyze_plan(self._fts_query.to_query_object()))
+        return "\n".join(results)
+
+    def _create_query_builders(self):
+        """Set up and configure the vector and FTS query builders."""
+        vector_query, fts_query = self._validate_query(
+            self._query, self._vector, self._text
+        )
+        self._fts_query = LanceFtsQueryBuilder(
+            self._table, fts_query, fts_columns=self._fts_columns
+        )
+        vector_query = self._query_to_vector(
+            self._table, vector_query, self._vector_column
+        )
+        self._vector_query = LanceVectorQueryBuilder(
+            self._table, vector_query, self._vector_column
+        )
+
+        # Apply common configurations
+        if self._limit:
+            self._vector_query.limit(self._limit)
+            self._fts_query.limit(self._limit)
+        if self._columns:
+            self._vector_query.select(self._columns)
+            self._fts_query.select(self._columns)
+        if self._where:
+            self._vector_query.where(self._where, self._postfilter)
+            self._fts_query.where(self._where, self._postfilter)
+        if self._with_row_id:
+            self._vector_query.with_row_id(True)
+            self._fts_query.with_row_id(True)
+        if self._phrase_query:
+            self._fts_query.phrase_query(True)
+        if self._distance_type:
+            self._vector_query.metric(self._distance_type)
+        if self._minimum_nprobes:
+            self._vector_query.minimum_nprobes(self._minimum_nprobes)
+        if self._maximum_nprobes is not None:
+            self._vector_query.maximum_nprobes(self._maximum_nprobes)
+        if self._refine_factor:
+            self._vector_query.refine_factor(self._refine_factor)
+        if self._ef:
+            self._vector_query.ef(self._ef)
+        if self._bypass_vector_index:
+            self._vector_query.bypass_vector_index()
+        if self._lower_bound or self._upper_bound:
+            self._vector_query.distance_range(
+                lower_bound=self._lower_bound, upper_bound=self._upper_bound
+            )
+
+        if self._reranker is None:
+            self._reranker = RRFReranker()
+
 
 class AsyncQueryBase(object):
     def __init__(self, inner: Union[LanceQuery, LanceVectorQuery]):
@@ -2439,7 +2577,7 @@ class AsyncQuery(AsyncQueryBase):
                 self._inner.nearest_to_text({"query": query, "columns": columns})
             )
         # FullTextQuery object
-        return AsyncFTSQuery(self._inner.nearest_to_text({"query": query.to_dict()}))
+        return AsyncFTSQuery(self._inner.nearest_to_text({"query": query}))
 
 
 class AsyncFTSQuery(AsyncQueryBase):
@@ -2587,6 +2725,34 @@ class AsyncVectorQueryBase:
         self._inner.nprobes(nprobes)
         return self
 
+    def minimum_nprobes(self, minimum_nprobes: int) -> Self:
+        """Set the minimum number of probes to use.
+
+        See `nprobes` for more details.
+
+        These partitions will be searched on every indexed vector query and will
+        increase recall at the expense of latency.
+        """
+        self._inner.minimum_nprobes(minimum_nprobes)
+        return self
+
+    def maximum_nprobes(self, maximum_nprobes: int) -> Self:
+        """Set the maximum number of probes to use.
+
+        See `nprobes` for more details.
+
+        If this value is greater than `minimum_nprobes` then the excess partitions
+        will be searched only if we have not found enough results.
+
+        This can be useful when there is a narrow filter to allow these queries to
+        spend more time searching and avoid potential false negatives.
+
+        If this value is 0 then no limit will be applied and all partitions could be
+        searched if needed to satisfy the limit.
+        """
+        self._inner.maximum_nprobes(maximum_nprobes)
+        return self
+
     def distance_range(
         self, lower_bound: Optional[float] = None, upper_bound: Optional[float] = None
     ) -> Self:
@@ -2761,7 +2927,7 @@ class AsyncVectorQuery(AsyncQueryBase, AsyncVectorQueryBase):
                 self._inner.nearest_to_text({"query": query, "columns": columns})
             )
         # FullTextQuery object
-        return AsyncHybridQuery(self._inner.nearest_to_text({"query": query.to_dict()}))
+        return AsyncHybridQuery(self._inner.nearest_to_text({"query": query}))
 
     async def to_batches(
         self,

python/python/lancedb/remote/table.py

@@ -7,7 +7,16 @@ from functools import cached_property
 from typing import Dict, Iterable, List, Optional, Union, Literal
 import warnings
 
-from lancedb._lancedb import IndexConfig
+from lancedb._lancedb import (
+    AddColumnsResult,
+    AddResult,
+    AlterColumnsResult,
+    DeleteResult,
+    DropColumnsResult,
+    IndexConfig,
+    MergeResult,
+    UpdateResult,
+)
 from lancedb.embeddings.base import EmbeddingFunctionConfig
 from lancedb.index import FTS, BTree, Bitmap, HnswPq, HnswSq, IvfFlat, IvfPq, LabelList
 from lancedb.remote.db import LOOP
@@ -18,7 +27,7 @@ from lancedb.merge import LanceMergeInsertBuilder
 from lancedb.embeddings import EmbeddingFunctionRegistry
 
 from ..query import LanceVectorQueryBuilder, LanceQueryBuilder
-from ..table import AsyncTable, IndexStatistics, Query, Table
+from ..table import AsyncTable, IndexStatistics, Query, Table, Tags
 
 
 class RemoteTable(Table):
@@ -38,9 +47,6 @@ class RemoteTable(Table):
     def __repr__(self) -> str:
         return f"RemoteTable({self.db_name}.{self.name})"
 
-    def __len__(self) -> int:
-        self.count_rows(None)
-
     @property
     def schema(self) -> pa.Schema:
         """The [Arrow Schema](https://arrow.apache.org/docs/python/api/datatypes.html#)
@@ -54,6 +60,10 @@ class RemoteTable(Table):
         """Get the current version of the table"""
         return LOOP.run(self._table.version())
 
+    @property
+    def tags(self) -> Tags:
+        return Tags(self._table)
+
     @cached_property
     def embedding_functions(self) -> Dict[str, EmbeddingFunctionConfig]:
         """
@@ -81,13 +91,13 @@ class RemoteTable(Table):
         """to_pandas() is not yet supported on LanceDB cloud."""
         return NotImplementedError("to_pandas() is not yet supported on LanceDB cloud.")
 
-    def checkout(self, version: int):
+    def checkout(self, version: Union[int, str]):
         return LOOP.run(self._table.checkout(version))
 
     def checkout_latest(self):
         return LOOP.run(self._table.checkout_latest())
 
-    def restore(self, version: Optional[int] = None):
+    def restore(self, version: Optional[Union[int, str]] = None):
         return LOOP.run(self._table.restore(version))
 
     def list_indices(self) -> Iterable[IndexConfig]:
@@ -104,6 +114,7 @@ class RemoteTable(Table):
         index_type: Literal["BTREE", "BITMAP", "LABEL_LIST", "scalar"] = "scalar",
         *,
         replace: bool = False,
+        wait_timeout: timedelta = None,
     ):
         """Creates a scalar index
         Parameters
@@ -126,22 +137,27 @@ class RemoteTable(Table):
         else:
             raise ValueError(f"Unknown index type: {index_type}")
 
-        LOOP.run(self._table.create_index(column, config=config, replace=replace))
+        LOOP.run(
+            self._table.create_index(
+                column, config=config, replace=replace, wait_timeout=wait_timeout
+            )
+        )
 
     def create_fts_index(
         self,
         column: str,
         *,
         replace: bool = False,
-        with_position: bool = True,
+        wait_timeout: timedelta = None,
+        with_position: bool = False,
         # tokenizer configs:
         base_tokenizer: str = "simple",
         language: str = "English",
         max_token_length: Optional[int] = 40,
         lower_case: bool = True,
-        stem: bool = False,
+        stem: bool = True,
-        remove_stop_words: bool = False,
+        remove_stop_words: bool = True,
-        ascii_folding: bool = False,
+        ascii_folding: bool = True,
     ):
         config = FTS(
             with_position=with_position,
@@ -153,7 +169,11 @@ class RemoteTable(Table):
             remove_stop_words=remove_stop_words,
             ascii_folding=ascii_folding,
         )
-        LOOP.run(self._table.create_index(column, config=config, replace=replace))
+        LOOP.run(
+            self._table.create_index(
+                column, config=config, replace=replace, wait_timeout=wait_timeout
+            )
+        )
 
     def create_index(
         self,
@@ -165,6 +185,7 @@ class RemoteTable(Table):
         replace: Optional[bool] = None,
         accelerator: Optional[str] = None,
         index_type="vector",
+        wait_timeout: Optional[timedelta] = None,
     ):
         """Create an index on the table.
         Currently, the only parameters that matter are
@@ -236,7 +257,11 @@ class RemoteTable(Table):
                 " 'IVF_FLAT', 'IVF_PQ', 'IVF_HNSW_PQ', 'IVF_HNSW_SQ'"
             )
 
-        LOOP.run(self._table.create_index(vector_column_name, config=config))
+        LOOP.run(
+            self._table.create_index(
+                vector_column_name, config=config, wait_timeout=wait_timeout
+            )
+        )
 
     def add(
         self,
@@ -244,7 +269,7 @@ class RemoteTable(Table):
         mode: str = "append",
         on_bad_vectors: str = "error",
         fill_value: float = 0.0,
-    ) -> int:
+    ) -> AddResult:
         """Add more data to the [Table](Table). It has the same API signature as
         the OSS version.
 
@@ -267,8 +292,12 @@ class RemoteTable(Table):
         fill_value: float, default 0.
             The value to use when filling vectors. Only used if on_bad_vectors="fill".
 
+        Returns
+        -------
+        AddResult
+            An object containing the new version number of the table after adding data.
         """
-        LOOP.run(
+        return LOOP.run(
             self._table.add(
                 data, mode=mode, on_bad_vectors=on_bad_vectors, fill_value=fill_value
             )
@@ -394,10 +423,12 @@ class RemoteTable(Table):
         new_data: DATA,
         on_bad_vectors: str,
         fill_value: float,
-    ):
+    ) -> MergeResult:
-        LOOP.run(self._table._do_merge(merge, new_data, on_bad_vectors, fill_value))
+        return LOOP.run(
+            self._table._do_merge(merge, new_data, on_bad_vectors, fill_value)
+        )
 
-    def delete(self, predicate: str):
+    def delete(self, predicate: str) -> DeleteResult:
         """Delete rows from the table.
 
         This can be used to delete a single row, many rows, all rows, or
@@ -412,6 +443,11 @@ class RemoteTable(Table):
 
             The filter must not be empty, or it will error.
 
+        Returns
+        -------
+        DeleteResult
+            An object containing the new version number of the table after deletion.
+
         Examples
         --------
         >>> import lancedb
@@ -444,7 +480,7 @@ class RemoteTable(Table):
            x      vector  _distance # doctest: +SKIP
         0  2  [3.0, 4.0]       85.0 # doctest: +SKIP
         """
-        LOOP.run(self._table.delete(predicate))
+        return LOOP.run(self._table.delete(predicate))
 
     def update(
         self,
@@ -452,7 +488,7 @@ class RemoteTable(Table):
         values: Optional[dict] = None,
         *,
         values_sql: Optional[Dict[str, str]] = None,
-    ):
+    ) -> UpdateResult:
         """
         This can be used to update zero to all rows depending on how many
         rows match the where clause.
@@ -470,6 +506,12 @@ class RemoteTable(Table):
             reference existing columns. For example, {"x": "x + 1"} will increment
             the x column by 1.
 
+        Returns
+        -------
+        UpdateResult
+            - rows_updated: The number of rows that were updated
+            - version: The new version number of the table after the update
+
         Examples
         --------
         >>> import lancedb
@@ -494,7 +536,7 @@ class RemoteTable(Table):
         2  2  [10.0, 10.0] # doctest: +SKIP
 
         """
-        LOOP.run(
+        return LOOP.run(
             self._table.update(where=where, updates=values, updates_sql=values_sql)
         )
 
@@ -542,18 +584,28 @@ class RemoteTable(Table):
     def count_rows(self, filter: Optional[str] = None) -> int:
         return LOOP.run(self._table.count_rows(filter))
 
-    def add_columns(self, transforms: Dict[str, str]):
+    def add_columns(self, transforms: Dict[str, str]) -> AddColumnsResult:
         return LOOP.run(self._table.add_columns(transforms))
 
-    def alter_columns(self, *alterations: Iterable[Dict[str, str]]):
+    def alter_columns(
+        self, *alterations: Iterable[Dict[str, str]]
+    ) -> AlterColumnsResult:
         return LOOP.run(self._table.alter_columns(*alterations))
 
-    def drop_columns(self, columns: Iterable[str]):
+    def drop_columns(self, columns: Iterable[str]) -> DropColumnsResult:
         return LOOP.run(self._table.drop_columns(columns))
 
     def drop_index(self, index_name: str):
         return LOOP.run(self._table.drop_index(index_name))
 
+    def wait_for_index(
+        self, index_names: Iterable[str], timeout: timedelta = timedelta(seconds=300)
+    ):
+        return LOOP.run(self._table.wait_for_index(index_names, timeout))
+
+    def stats(self):
+        return LOOP.run(self._table.stats())
+
     def uses_v2_manifest_paths(self) -> bool:
         raise NotImplementedError(
             "uses_v2_manifest_paths() is not supported on the LanceDB Cloud"

python/python/tests/docs/test_guide_tables.py

@@ -25,6 +25,10 @@ import numpy as np
 from lancedb.pydantic import Vector, LanceModel
 
 # --8<-- [end:import-lancedb-pydantic]
+# --8<-- [start:import-session-context]
+from datafusion import SessionContext
+
+# --8<-- [end:import-session-context]
 # --8<-- [start:import-datetime]
 from datetime import timedelta
 
@@ -33,6 +37,10 @@ from datetime import timedelta
 from lancedb.embeddings import get_registry
 
 # --8<-- [end:import-embeddings]
+# --8<-- [start:import-ffi-dataset]
+from lance import FFILanceTableProvider
+
+# --8<-- [end:import-ffi-dataset]
 # --8<-- [start:import-pydantic-basemodel]
 from pydantic import BaseModel
 
@@ -341,6 +349,27 @@ def test_table_with_embedding():
     # --8<-- [end:create_table_with_embedding]
 
 
+def test_sql_query():
+    db = lancedb.connect("data/sample-lancedb")
+    data = [
+        {"vector": [1.1, 1.2], "lat": 45.5, "long": -122.7},
+        {"vector": [0.2, 1.8], "lat": 40.1, "long": -74.1},
+    ]
+    table = db.create_table("lance_table", data)
+
+    # --8<-- [start:lance_sql_basic]
+    ctx = SessionContext()
+    ffi_lance_table = FFILanceTableProvider(
+        table.to_lance(), with_row_id=False, with_row_addr=False
+    )
+
+    ctx.register_table_provider("ffi_lance_table", ffi_lance_table)
+    ctx.table("ffi_lance_table")
+
+    ctx.sql("SELECT vector FROM ffi_lance_table")
+    # --8<-- [end:lance_sql_basic]
+
+
 @pytest.mark.skip
 async def test_table_with_embedding_async():
     async_db = await lancedb.connect_async("data/sample-lancedb")

python/python/tests/docs/test_merge_insert.py

@@ -18,15 +18,19 @@ def test_upsert(mem_db):
         {"id": 1, "name": "Bobby"},
         {"id": 2, "name": "Charlie"},
     ]
-    (
+    res = (
         table.merge_insert("id")
         .when_matched_update_all()
         .when_not_matched_insert_all()
         .execute(new_users)
     )
     table.count_rows()  # 3
+    res  # {'num_inserted_rows': 1, 'num_updated_rows': 1, 'num_deleted_rows': 0}
     # --8<-- [end:upsert_basic]
     assert table.count_rows() == 3
+    assert res.num_inserted_rows == 1
+    assert res.num_deleted_rows == 0
+    assert res.num_updated_rows == 1
 
 
 @pytest.mark.asyncio
@@ -44,15 +48,22 @@ async def test_upsert_async(mem_db_async):
         {"id": 1, "name": "Bobby"},
         {"id": 2, "name": "Charlie"},
     ]
-    await (
+    res = await (
         table.merge_insert("id")
         .when_matched_update_all()
         .when_not_matched_insert_all()
         .execute(new_users)
     )
     await table.count_rows()  # 3
+    res
+    # MergeResult(version=2, num_updated_rows=1,
+    # num_inserted_rows=1, num_deleted_rows=0)
     # --8<-- [end:upsert_basic_async]
     assert await table.count_rows() == 3
+    assert res.version == 2
+    assert res.num_inserted_rows == 1
+    assert res.num_deleted_rows == 0
+    assert res.num_updated_rows == 1
 
 
 def test_insert_if_not_exists(mem_db):
@@ -69,10 +80,19 @@ def test_insert_if_not_exists(mem_db):
         {"domain": "google.com", "name": "Google"},
         {"domain": "facebook.com", "name": "Facebook"},
     ]
-    (table.merge_insert("domain").when_not_matched_insert_all().execute(new_domains))
+    res = (
+        table.merge_insert("domain").when_not_matched_insert_all().execute(new_domains)
+    )
     table.count_rows()  # 3
+    res
+    # MergeResult(version=2, num_updated_rows=0,
+    # num_inserted_rows=1, num_deleted_rows=0)
     # --8<-- [end:insert_if_not_exists]
     assert table.count_rows() == 3
+    assert res.version == 2
+    assert res.num_inserted_rows == 1
+    assert res.num_deleted_rows == 0
+    assert res.num_updated_rows == 0
 
 
 @pytest.mark.asyncio
@@ -90,12 +110,19 @@ async def test_insert_if_not_exists_async(mem_db_async):
         {"domain": "google.com", "name": "Google"},
         {"domain": "facebook.com", "name": "Facebook"},
     ]
-    await (
+    res = await (
         table.merge_insert("domain").when_not_matched_insert_all().execute(new_domains)
     )
     await table.count_rows()  # 3
-    # --8<-- [end:insert_if_not_exists_async]
+    res
+    # MergeResult(version=2, num_updated_rows=0,
+    # num_inserted_rows=1, num_deleted_rows=0)
+    # --8<-- [end:insert_if_not_exists]
     assert await table.count_rows() == 3
+    assert res.version == 2
+    assert res.num_inserted_rows == 1
+    assert res.num_deleted_rows == 0
+    assert res.num_updated_rows == 0
 
 
 def test_replace_range(mem_db):
@@ -113,7 +140,7 @@ def test_replace_range(mem_db):
     new_chunks = [
         {"doc_id": 1, "chunk_id": 0, "text": "Baz"},
     ]
-    (
+    res = (
         table.merge_insert(["doc_id", "chunk_id"])
         .when_matched_update_all()
         .when_not_matched_insert_all()
@@ -121,8 +148,15 @@ def test_replace_range(mem_db):
         .execute(new_chunks)
     )
     table.count_rows("doc_id = 1")  # 1
-    # --8<-- [end:replace_range]
+    res
+    # MergeResult(version=2, num_updated_rows=1,
+    # num_inserted_rows=0, num_deleted_rows=1)
+    # --8<-- [end:insert_if_not_exists]
     assert table.count_rows("doc_id = 1") == 1
+    assert res.version == 2
+    assert res.num_inserted_rows == 0
+    assert res.num_deleted_rows == 1
+    assert res.num_updated_rows == 1
 
 
 @pytest.mark.asyncio
@@ -141,7 +175,7 @@ async def test_replace_range_async(mem_db_async):
     new_chunks = [
         {"doc_id": 1, "chunk_id": 0, "text": "Baz"},
     ]
-    await (
+    res = await (
         table.merge_insert(["doc_id", "chunk_id"])
         .when_matched_update_all()
         .when_not_matched_insert_all()
@@ -149,5 +183,12 @@ async def test_replace_range_async(mem_db_async):
         .execute(new_chunks)
     )
     await table.count_rows("doc_id = 1")  # 1
-    # --8<-- [end:replace_range_async]
+    res
+    # MergeResult(version=2, num_updated_rows=1,
+    # num_inserted_rows=0, num_deleted_rows=1)
+    # --8<-- [end:insert_if_not_exists]
     assert await table.count_rows("doc_id = 1") == 1
+    assert res.version == 2
+    assert res.num_inserted_rows == 0
+    assert res.num_deleted_rows == 1
+    assert res.num_updated_rows == 1

python/python/tests/docs/test_search.py

@@ -6,7 +6,7 @@ import lancedb
 
 # --8<-- [end:import-lancedb]
 # --8<-- [start:import-numpy]
-from lancedb.query import BoostQuery, MatchQuery
+from lancedb.query import BooleanQuery, BoostQuery, MatchQuery, Occur
 import numpy as np
 import pyarrow as pa
 
@@ -156,6 +156,9 @@ async def test_vector_search_async():
     # --8<-- [end:search_result_async_as_list]
 
 
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
 def test_fts_fuzzy_query():
     uri = "data/fuzzy-example"
     db = lancedb.connect(uri)
@@ -188,7 +191,19 @@ def test_fts_fuzzy_query():
         "food",  # 1 insertion
     }
 
+    results = table.search(
+        MatchQuery("foo", "text", fuzziness=1, prefix_length=3)
+    ).to_pandas()
+    assert len(results) == 2
+    assert set(results["text"].to_list()) == {
+        "foo",
+        "food",
+    }
 
+
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
 def test_fts_boost_query():
     uri = "data/boost-example"
     db = lancedb.connect(uri)
@@ -234,6 +249,63 @@ def test_fts_boost_query():
     )
 
 
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
+def test_fts_boolean_query(tmp_path):
+    uri = tmp_path / "boolean-example"
+    db = lancedb.connect(uri)
+    table = db.create_table(
+        "my_table_fts_boolean",
+        data=[
+            {"text": "The cat and dog are playing"},
+            {"text": "The cat is sleeping"},
+            {"text": "The dog is barking"},
+            {"text": "The dog chases the cat"},
+        ],
+        mode="overwrite",
+    )
+    table.create_fts_index("text", use_tantivy=False, replace=True)
+
+    # SHOULD
+    results = table.search(
+        MatchQuery("cat", "text") | MatchQuery("dog", "text")
+    ).to_pandas()
+    assert len(results) == 4
+    assert set(results["text"].to_list()) == {
+        "The cat and dog are playing",
+        "The cat is sleeping",
+        "The dog is barking",
+        "The dog chases the cat",
+    }
+    # MUST
+    results = table.search(
+        MatchQuery("cat", "text") & MatchQuery("dog", "text")
+    ).to_pandas()
+    assert len(results) == 2
+    assert set(results["text"].to_list()) == {
+        "The cat and dog are playing",
+        "The dog chases the cat",
+    }
+
+    # MUST NOT
+    results = table.search(
+        BooleanQuery(
+            [
+                (Occur.MUST, MatchQuery("cat", "text")),
+                (Occur.MUST_NOT, MatchQuery("dog", "text")),
+            ]
+        )
+    ).to_pandas()
+    assert len(results) == 1
+    assert set(results["text"].to_list()) == {
+        "The cat is sleeping",
+    }
+
+
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
 def test_fts_native():
     # --8<-- [start:basic_fts]
     uri = "data/sample-lancedb"
@@ -282,6 +354,9 @@ def test_fts_native():
     # --8<-- [end:fts_incremental_index]
 
 
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
 @pytest.mark.asyncio
 async def test_fts_native_async():
     # --8<-- [start:basic_fts_async]

python/python/tests/test_embeddings_slow.py

@@ -11,7 +11,7 @@ import pandas as pd
 import pyarrow as pa
 import pytest
 from lancedb.embeddings import get_registry
-from lancedb.pydantic import LanceModel, Vector
+from lancedb.pydantic import LanceModel, Vector, MultiVector
 import requests
 
 # These are integration tests for embedding functions.
@@ -575,3 +575,67 @@ def test_voyageai_multimodal_embedding_text_function():
 
     tbl.add(df)
     assert len(tbl.to_pandas()["vector"][0]) == voyageai.ndims()
+
+
+@pytest.mark.slow
+@pytest.mark.skipif(
+    importlib.util.find_spec("colpali_engine") is None,
+    reason="colpali_engine not installed",
+)
+def test_colpali(tmp_path):
+    import requests
+    from lancedb.pydantic import LanceModel
+
+    db = lancedb.connect(tmp_path)
+    registry = get_registry()
+    func = registry.get("colpali").create()
+
+    class MediaItems(LanceModel):
+        text: str
+        image_uri: str = func.SourceField()
+        image_bytes: bytes = func.SourceField()
+        image_vectors: MultiVector(func.ndims()) = (
+            func.VectorField()
+        )  # Multivector image embeddings
+
+    table = db.create_table("media", schema=MediaItems)
+
+    texts = [
+        "a cute cat playing with yarn",
+        "a puppy in a flower field",
+        "a red sports car on the highway",
+        "a vintage bicycle leaning against a wall",
+        "a plate of delicious pasta",
+        "fresh fruit salad in a bowl",
+    ]
+
+    uris = [
+        "http://farm1.staticflickr.com/53/167798175_7c7845bbbd_z.jpg",
+        "http://farm1.staticflickr.com/134/332220238_da527d8140_z.jpg",
+        "http://farm9.staticflickr.com/8387/8602747737_2e5c2a45d4_z.jpg",
+        "http://farm5.staticflickr.com/4092/5017326486_1f46057f5f_z.jpg",
+        "http://farm9.staticflickr.com/8216/8434969557_d37882c42d_z.jpg",
+        "http://farm6.staticflickr.com/5142/5835678453_4f3a4edb45_z.jpg",
+    ]
+
+    # Get images as bytes
+    image_bytes = [requests.get(uri).content for uri in uris]
+
+    table.add(
+        pd.DataFrame({"text": texts, "image_uri": uris, "image_bytes": image_bytes})
+    )
+
+    # Test text-to-image search
+    image_results = (
+        table.search("fluffy companion", vector_column_name="image_vectors")
+        .limit(1)
+        .to_pydantic(MediaItems)[0]
+    )
+    assert "cat" in image_results.text.lower() or "puppy" in image_results.text.lower()
+
+    # Verify multivector dimensions
+    first_row = table.to_arrow().to_pylist()[0]
+    assert len(first_row["image_vectors"]) > 1, "Should have multiple image vectors"
+    assert len(first_row["image_vectors"][0]) == func.ndims(), (
+        "Vector dimension mismatch"
+    )

python/python/tests/test_fts.py

@@ -215,6 +215,19 @@ def test_search_fts(table, use_tantivy):
         assert len(results) == 5
         assert len(results[0]) == 3  # id, text, _score
 
+        # Test boolean query
+        results = (
+            table.search(MatchQuery("puppy", "text") & MatchQuery("runs", "text"))
+            .select(["id", "text"])
+            .limit(5)
+            .to_list()
+        )
+        assert len(results) == 5
+        assert len(results[0]) == 3  # id, text, _score
+        for r in results:
+            assert "puppy" in r["text"]
+            assert "runs" in r["text"]
+
 
 @pytest.mark.asyncio
 async def test_fts_select_async(async_table):
@@ -287,7 +300,7 @@ def test_search_fts_phrase_query(table):
         assert False
     except Exception:
         pass
-    table.create_fts_index("text", use_tantivy=False, replace=True)
+    table.create_fts_index("text", use_tantivy=False, with_position=True, replace=True)
     results = table.search("puppy").limit(100).to_list()
     phrase_results = table.search('"puppy runs"').limit(100).to_list()
     assert len(results) > len(phrase_results)
@@ -312,7 +325,7 @@ async def test_search_fts_phrase_query_async(async_table):
         assert False
     except Exception:
         pass
-    await async_table.create_index("text", config=FTS())
+    await async_table.create_index("text", config=FTS(with_position=True))
     results = await async_table.query().nearest_to_text("puppy").limit(100).to_list()
     phrase_results = (
         await async_table.query().nearest_to_text('"puppy runs"').limit(100).to_list()
@@ -649,7 +662,7 @@ def test_fts_on_list(mem_db: DBConnection):
         }
     )
     table = mem_db.create_table("test", data=data)
-    table.create_fts_index("text", use_tantivy=False)
+    table.create_fts_index("text", use_tantivy=False, with_position=True)
 
     res = table.search("lance").limit(5).to_list()
     assert len(res) == 3

python/python/tests/test_hybrid_query.py

@@ -4,13 +4,32 @@
 import lancedb
 
 from lancedb.query import LanceHybridQueryBuilder
+from lancedb.rerankers.rrf import RRFReranker
 import pyarrow as pa
 import pyarrow.compute as pc
 import pytest
 import pytest_asyncio
 
 from lancedb.index import FTS
-from lancedb.table import AsyncTable
+from lancedb.table import AsyncTable, Table
+
+
+@pytest.fixture
+def sync_table(tmpdir_factory) -> Table:
+    tmp_path = str(tmpdir_factory.mktemp("data"))
+    db = lancedb.connect(tmp_path)
+    data = pa.table(
+        {
+            "text": pa.array(["a", "b", "cat", "dog"]),
+            "vector": pa.array(
+                [[0.1, 0.1], [2, 2], [-0.1, -0.1], [0.5, -0.5]],
+                type=pa.list_(pa.float32(), list_size=2),
+            ),
+        }
+    )
+    table = db.create_table("test", data)
+    table.create_fts_index("text", with_position=False, use_tantivy=False)
+    return table
 
 
 @pytest_asyncio.fixture
@@ -102,6 +121,42 @@ async def test_async_hybrid_query_default_limit(table: AsyncTable):
     assert texts.count("a") == 1
 
 
+def test_hybrid_query_distance_range(sync_table: Table):
+    reranker = RRFReranker(return_score="all")
+    result = (
+        sync_table.search(query_type="hybrid")
+        .vector([0.0, 0.4])
+        .text("cat and dog")
+        .distance_range(lower_bound=0.2, upper_bound=0.5)
+        .rerank(reranker)
+        .limit(2)
+        .to_arrow()
+    )
+    assert len(result) == 2
+    print(result)
+    for dist in result["_distance"]:
+        if dist.is_valid:
+            assert 0.2 <= dist.as_py() <= 0.5
+
+
+@pytest.mark.asyncio
+async def test_hybrid_query_distance_range_async(table: AsyncTable):
+    reranker = RRFReranker(return_score="all")
+    result = await (
+        table.query()
+        .nearest_to([0.0, 0.4])
+        .nearest_to_text("cat and dog")
+        .distance_range(lower_bound=0.2, upper_bound=0.5)
+        .rerank(reranker)
+        .limit(2)
+        .to_arrow()
+    )
+    assert len(result) == 2
+    for dist in result["_distance"]:
+        if dist.is_valid:
+            assert 0.2 <= dist.as_py() <= 0.5
+
+
 @pytest.mark.asyncio
 async def test_explain_plan(table: AsyncTable):
     plan = await (

python/python/tests/test_index.py

@@ -8,7 +8,7 @@ import pyarrow as pa
 import pytest
 import pytest_asyncio
 from lancedb import AsyncConnection, AsyncTable, connect_async
-from lancedb.index import BTree, IvfFlat, IvfPq, Bitmap, LabelList, HnswPq, HnswSq
+from lancedb.index import BTree, IvfFlat, IvfPq, Bitmap, LabelList, HnswPq, HnswSq, FTS
 
 
 @pytest_asyncio.fixture
@@ -119,6 +119,18 @@ async def test_create_label_list_index(some_table: AsyncTable):
     assert str(indices) == '[Index(LabelList, columns=["tags"], name="tags_idx")]'
 
 
+@pytest.mark.asyncio
+async def test_full_text_search_index(some_table: AsyncTable):
+    await some_table.create_index("tags", config=FTS(with_position=False))
+    indices = await some_table.list_indices()
+    assert str(indices) == '[Index(FTS, columns=["tags"], name="tags_idx")]'
+
+    await some_table.prewarm_index("tags_idx")
+
+    res = await (await some_table.search("tag0")).to_arrow()
+    assert res.num_rows > 0
+
+
 @pytest.mark.asyncio
 async def test_create_vector_index(some_table: AsyncTable):
     # Can create