Bump version: 0.30.1 → 0.30.2-beta.0

Revert "feat: allow passing azure client/tenant ID through remote SDK" (#3185 )
Reverts lancedb/lancedb#3102
2026-03-26 02:20:40 +00:00 · 2026-03-25 03:22:55 +00:00 · 2026-03-24 20:17:40 -07:00 · 2026-03-23 16:14:13 -07:00 · 2026-03-23 13:45:44 -07:00 · 2026-03-23 15:30:33 -05:00
56 changed files with 2438 additions and 397 deletions
--- a/.bumpversion.toml
+++ b/.bumpversion.toml
@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.27.0-beta.5"
+current_version = "0.27.1"
 parse = """(?x)
    (?P<major>0|[1-9]\\d*)\\.
    (?P<minor>0|[1-9]\\d*)\\.
--- a/.github/workflows/nodejs.yml
+++ b/.github/workflows/nodejs.yml
@@ -7,6 +7,7 @@ on:
  pull_request:
    paths:
      - Cargo.toml
+      - Cargo.lock
      - nodejs/**
      - rust/**
      - docs/src/js/**
--- a/.github/workflows/npm-publish.yml
+++ b/.github/workflows/npm-publish.yml
@@ -19,6 +19,7 @@ on:
    paths:
      - .github/workflows/npm-publish.yml
      - Cargo.toml # Change in dependency frequently breaks builds
+      - Cargo.lock

 concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
@@ -124,7 +125,12 @@ jobs:
            pre_build: |-
              set -e &&
              apt-get update &&
-              apt-get install -y protobuf-compiler pkg-config
+              apt-get install -y protobuf-compiler pkg-config &&
+              # The base image (manylinux2014-cross) sets TARGET_CC to the old
+              # GCC 4.8 cross-compiler. aws-lc-sys checks TARGET_CC before CC,
+              # so it picks up GCC even though the napi-rs image sets CC=clang.
+              # Override to use the image's clang-18 which supports -fuse-ld=lld.
+              export TARGET_CC=clang TARGET_CXX=clang++
          - target: x86_64-unknown-linux-musl
            # This one seems to need some extra memory
            host: ubuntu-2404-8x-x64
@@ -144,9 +150,10 @@ jobs:
              set -e &&
              apt-get update &&
              apt-get install -y protobuf-compiler pkg-config &&
-              # https://github.com/aws/aws-lc-rs/issues/737#issuecomment-2725918627
-              ln -s /usr/aarch64-unknown-linux-gnu/lib/gcc/aarch64-unknown-linux-gnu/4.8.5/crtbeginS.o /usr/aarch64-unknown-linux-gnu/aarch64-unknown-linux-gnu/sysroot/usr/lib/crtbeginS.o &&
-              ln -s /usr/aarch64-unknown-linux-gnu/lib/gcc /usr/aarch64-unknown-linux-gnu/aarch64-unknown-linux-gnu/sysroot/usr/lib/gcc &&
+              export TARGET_CC=clang TARGET_CXX=clang++ &&
+              # The manylinux2014 sysroot has glibc 2.17 headers which lack
+              # AT_HWCAP2 (added in Linux 3.17). Define it for aws-lc-sys.
+              export CFLAGS="$CFLAGS -DAT_HWCAP2=26" &&
              rustup target add aarch64-unknown-linux-gnu
          - target: aarch64-unknown-linux-musl
            host: ubuntu-2404-8x-x64
@@ -266,7 +273,7 @@ jobs:
          - target: x86_64-unknown-linux-gnu
            host: ubuntu-latest
          - target: aarch64-unknown-linux-gnu
-            host: buildjet-16vcpu-ubuntu-2204-arm
+            host: ubuntu-2404-8x-arm64
        node:
          - '20'
    runs-on: ${{ matrix.settings.host }}
--- a/.github/workflows/pypi-publish.yml
+++ b/.github/workflows/pypi-publish.yml
@@ -9,6 +9,7 @@ on:
    paths:
      - .github/workflows/pypi-publish.yml
      - Cargo.toml # Change in dependency frequently breaks builds
+      - Cargo.lock

 env:
  PIP_EXTRA_INDEX_URL: "https://pypi.fury.io/lance-format/ https://pypi.fury.io/lancedb/"
--- a/.github/workflows/python.yml
+++ b/.github/workflows/python.yml
@@ -7,6 +7,7 @@ on:
  pull_request:
    paths:
      - Cargo.toml
+      - Cargo.lock
      - python/**
      - rust/**
      - .github/workflows/python.yml
--- a/.github/workflows/rust.yml
+++ b/.github/workflows/rust.yml
@@ -7,6 +7,7 @@ on:
  pull_request:
    paths:
      - Cargo.toml
+      - Cargo.lock
      - rust/**
      - .github/workflows/rust.yml

@@ -206,14 +207,14 @@ jobs:
      - name: Downgrade  dependencies
        # These packages have newer requirements for MSRV
        run: |
-          cargo update -p aws-sdk-bedrockruntime --precise 1.64.0
-          cargo update -p aws-sdk-dynamodb --precise 1.55.0
-          cargo update -p aws-config --precise 1.5.10
-          cargo update -p aws-sdk-kms --precise 1.51.0
-          cargo update -p aws-sdk-s3 --precise 1.65.0
-          cargo update -p aws-sdk-sso --precise 1.50.0
-          cargo update -p aws-sdk-ssooidc --precise 1.51.0
-          cargo update -p aws-sdk-sts --precise 1.51.0
+          cargo update -p aws-sdk-bedrockruntime --precise 1.77.0
+          cargo update -p aws-sdk-dynamodb --precise 1.68.0
+          cargo update -p aws-config --precise 1.6.0
+          cargo update -p aws-sdk-kms --precise 1.63.0
+          cargo update -p aws-sdk-s3 --precise 1.79.0
+          cargo update -p aws-sdk-sso --precise 1.62.0
+          cargo update -p aws-sdk-ssooidc --precise 1.63.0
+          cargo update -p aws-sdk-sts --precise 1.63.0
          cargo update -p home --precise 0.5.9
      - name: cargo +${{ matrix.msrv }} check
        env:
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -3070,8 +3070,9 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"

 [[package]]
 name = "fsst"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a32ddfc5478379cd1782bdd9d7d1411063f563e5b338fc73bafe5916451a5b9d"
 dependencies = [
 "arrow-array",
 "rand 0.9.2",
@@ -3852,7 +3853,7 @@ dependencies = [
 "libc",
 "percent-encoding",
 "pin-project-lite",
- "socket2 0.5.10",
+ "socket2 0.6.0",
 "system-configuration",
 "tokio",
 "tower-service",
@@ -4241,8 +4242,9 @@ dependencies = [

 [[package]]
 name = "lance"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "95c5ce428fda0721f5c48bfde17a1921c4da2d2142b2f46a16c89abf5fce8003"
 dependencies = [
 "arrow",
 "arrow-arith",
@@ -4308,8 +4310,9 @@ dependencies = [

 [[package]]
 name = "lance-arrow"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c9fdaf99863fa0d631e422881e88be4837d8b82f36a87143d723a9d285acec4b"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4329,8 +4332,9 @@ dependencies = [

 [[package]]
 name = "lance-bitpacking"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "866b1634d38d94e8ab86fbcf238ac82dc8a5f72a4a6a90525f29899772e7cc7f"
 dependencies = [
 "arrayref",
 "paste",
@@ -4339,8 +4343,9 @@ dependencies = [

 [[package]]
 name = "lance-core"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "977c29f4e48c201c2806fe6ae117b65d0287eda236acd07357b556a54b0d5c5a"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4377,8 +4382,9 @@ dependencies = [

 [[package]]
 name = "lance-datafusion"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ccc72695473f4207df4c6df3b347a63e84c32c0bc36bf42a7d86e8a7c0c67e2"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -4408,8 +4414,9 @@ dependencies = [

 [[package]]
 name = "lance-datagen"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8fe84d76944acd834ded14d7562663af995556e0c6594f4b4ac69b0183f99c1a"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -4427,8 +4434,9 @@ dependencies = [

 [[package]]
 name = "lance-encoding"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "be1007242188e5d53c98717e7f2cb340dc80eb9c94c2b935587598919b3a36bd"
 dependencies = [
 "arrow-arith",
 "arrow-array",
@@ -4465,8 +4473,9 @@ dependencies = [

 [[package]]
 name = "lance-file"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f80088e418941f39cf5599d166ae1a6ef498cc2d967652a0692477d4871a9277"
 dependencies = [
 "arrow-arith",
 "arrow-array",
@@ -4498,8 +4507,9 @@ dependencies = [

 [[package]]
 name = "lance-index"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e0011daf1ddde99becffd2ae235ad324576736a526c54ffbc4d7e583872f1215"
 dependencies = [
 "arrow",
 "arrow-arith",
@@ -4562,8 +4572,9 @@ dependencies = [

 [[package]]
 name = "lance-io"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cfa8a74e93753d19a27ce3adaeb99e31227df13ad5926dd43572be76b43dd284"
 dependencies = [
 "arrow",
 "arrow-arith",
@@ -4604,8 +4615,9 @@ dependencies = [

 [[package]]
 name = "lance-linalg"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6e2d8da8f6b8dd37ab3b8199896ee265817f86232e3727c0b0eeb3c9093b64d9"
 dependencies = [
 "arrow-array",
 "arrow-buffer",
@@ -4621,8 +4633,9 @@ dependencies = [

 [[package]]
 name = "lance-namespace"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f176e427d9c35938d8a7097876114bc35dfd280b06077779753f2effe3e86aab"
 dependencies = [
 "arrow",
 "async-trait",
@@ -4634,8 +4647,9 @@ dependencies = [

 [[package]]
 name = "lance-namespace-impls"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "663c32086ecfab311acb0813c65a4bb352a5b648ccf8b513c24697ce8d412039"
 dependencies = [
 "arrow",
 "arrow-ipc",
@@ -4679,8 +4693,9 @@ dependencies = [

 [[package]]
 name = "lance-table"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "aa189b3081481a97b64cf1161297947a63b8adb941b1950989d0269858703a43"
 dependencies = [
 "arrow",
 "arrow-array",
@@ -4719,8 +4734,9 @@ dependencies = [

 [[package]]
 name = "lance-testing"
-version = "3.0.0-rc.3"
-source = "git+https://github.com/lance-format/lance.git?tag=v3.0.0-rc.3#de393a26a068dd297929ca7d798e43dc31c57337"
+version = "3.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "79a6f4ab0788ee82893bac5de4ff0d0d88bba96de87db4b6e18b1883616d4dbe"
 dependencies = [
 "arrow-array",
 "arrow-schema",
@@ -4731,7 +4747,7 @@ dependencies = [

 [[package]]
 name = "lancedb"
-version = "0.27.0-beta.5"
+version = "0.27.1"
 dependencies = [
 "ahash",
 "anyhow",
@@ -4813,7 +4829,7 @@ dependencies = [

 [[package]]
 name = "lancedb-nodejs"
-version = "0.27.0-beta.5"
+version = "0.27.1"
 dependencies = [
 "arrow-array",
 "arrow-ipc",
@@ -4833,7 +4849,7 @@ dependencies = [

 [[package]]
 name = "lancedb-python"
-version = "0.30.0-beta.5"
+version = "0.30.1"
 dependencies = [
 "arrow",
 "async-trait",
@@ -4845,6 +4861,7 @@ dependencies = [
 "lance-namespace",
 "lance-namespace-impls",
 "lancedb",
+ "log",
 "pin-project",
 "pyo3",
 "pyo3-async-runtimes",
@@ -6443,7 +6460,7 @@ version = "0.14.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "343d3bd7056eda839b03204e68deff7d1b13aba7af2b2fd16890697274262ee7"
 dependencies = [
- "heck 0.4.1",
+ "heck 0.5.0",
 "itertools 0.14.0",
 "log",
 "multimap",
@@ -6632,7 +6649,7 @@ dependencies = [
 "quinn-udp",
 "rustc-hash",
 "rustls 0.23.31",
- "socket2 0.5.10",
+ "socket2 0.6.0",
 "thiserror 2.0.17",
 "tokio",
 "tracing",
@@ -6669,7 +6686,7 @@ dependencies = [
 "cfg_aliases",
 "libc",
 "once_cell",
- "socket2 0.5.10",
+ "socket2 0.6.0",
 "tracing",
 "windows-sys 0.60.2",
 ]
@@ -7735,7 +7752,7 @@ version = "0.8.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c1c97747dbf44bb1ca44a561ece23508e99cb592e862f22222dcf42f51d1e451"
 dependencies = [
- "heck 0.4.1",
+ "heck 0.5.0",
 "proc-macro2",
 "quote",
 "syn 2.0.114",
@@ -7747,7 +7764,7 @@ version = "0.9.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "54254b8531cafa275c5e096f62d48c81435d1015405a91198ddb11e967301d40"
 dependencies = [
- "heck 0.4.1",
+ "heck 0.5.0",
 "proc-macro2",
 "quote",
 "syn 2.0.114",
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -15,20 +15,20 @@ categories = ["database-implementations"]
 rust-version = "1.91.0"

 [workspace.dependencies]
-lance = { "version" = "=3.0.0-rc.3", default-features = false, "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-core = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-datagen = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-file = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-io = { "version" = "=3.0.0-rc.3", default-features = false, "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-index = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-linalg = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-namespace-impls = { "version" = "=3.0.0-rc.3", default-features = false, "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-table = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-testing = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-datafusion = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-encoding = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
-lance-arrow = { "version" = "=3.0.0-rc.3", "tag" = "v3.0.0-rc.3", "git" = "https://github.com/lance-format/lance.git" }
+lance = { version = "=3.0.1", default-features = false }
+lance-core = { version = "=3.0.1" }
+lance-datagen = { version = "=3.0.1" }
+lance-file = { version = "=3.0.1" }
+lance-io = { version = "=3.0.1", default-features = false }
+lance-index = { version = "=3.0.1" }
+lance-linalg = { version = "=3.0.1" }
+lance-namespace = { version = "=3.0.1" }
+lance-namespace-impls = { version = "=3.0.1", default-features = false }
+lance-table = { version = "=3.0.1" }
+lance-testing = { version = "=3.0.1" }
+lance-datafusion = { version = "=3.0.1" }
+lance-encoding = { version = "=3.0.1" }
+lance-arrow = { version = "=3.0.1" }
 ahash = "0.8"
 # Note that this one does not include pyarrow
 arrow = { version = "57.2", optional = false }
--- a/ci/check_lance_release.py
+++ b/ci/check_lance_release.py
@@ -3,6 +3,7 @@
 from __future__ import annotations

 import argparse
+import functools
 import json
 import os
 import re
@@ -26,6 +27,7 @@ SEMVER_RE = re.compile(
 )


+@functools.total_ordering
@dataclass(frozen=True)
 class SemVer:
    major: int
@@ -156,7 +158,9 @@ def read_current_version(repo_root: Path) -> str:


 def determine_latest_tag(tags: Iterable[TagInfo]) -> TagInfo:
-    return max(tags, key=lambda tag: tag.semver)
+    # Stable releases (no prerelease) are always preferred over pre-releases.
+    # Within each group, standard semver ordering applies.
+    return max(tags, key=lambda tag: (not tag.semver.prerelease, tag.semver))


 def write_outputs(args: argparse.Namespace, payload: dict) -> None:
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -1,7 +1,7 @@
 version: "3.9"
 services:
  localstack:
-    image: localstack/localstack:3.3
+    image: localstack/localstack:4.0
    ports:
      - 4566:4566
    environment:
--- a/docs/requirements.txt
+++ b/docs/requirements.txt
@@ -1,8 +1,8 @@
-mkdocs==1.5.3
+mkdocs==1.6.1
 mkdocs-jupyter==0.24.1
-mkdocs-material==9.5.3
+mkdocs-material==9.6.23
 mkdocs-autorefs>=0.5,<=1.0
-mkdocstrings[python]==0.25.2
+mkdocstrings[python]>=0.24,<1.0
 griffe>=0.40,<1.0
 mkdocs-render-swagger-plugin>=0.1.0
 pydantic>=2.0,<3.0
--- a/docs/src/java/java.md
+++ b/docs/src/java/java.md
@@ -14,7 +14,7 @@ Add the following dependency to your `pom.xml`:
 <dependency>
    <groupId>com.lancedb</groupId>
    <artifactId>lancedb-core</artifactId>
-    <version>0.27.0-beta.5</version>
+    <version>0.27.1</version>
 </dependency>
 ```

--- a/docs/src/js/classes/Table.md
+++ b/docs/src/js/classes/Table.md
@@ -71,11 +71,12 @@ Add new columns with defined values.

 #### Parameters

-* **newColumnTransforms**: [`AddColumnsSql`](../interfaces/AddColumnsSql.md)[]
-    pairs of column names and
-    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.
+* **newColumnTransforms**: `Field`&lt;`any`&gt; \| `Field`&lt;`any`&gt;[] \| `Schema`&lt;`any`&gt; \| [`AddColumnsSql`](../interfaces/AddColumnsSql.md)[]
+    Either:
+    - An array of objects with column names and SQL expressions to calculate values
+    - A single Arrow Field defining one column with its data type (column will be initialized with null values)
+    - An array of Arrow Fields defining columns with their data types (columns will be initialized with null values)
+    - An Arrow Schema defining columns with their data types (columns will be initialized with null values)

 #### Returns

@@ -484,19 +485,7 @@ Modeled after ``VACUUM`` in PostgreSQL.
 - Prune: Removes old versions of the dataset
 - Index: Optimizes the indices, adding new data to existing indices

- Experimental API
- ----------------
-
- The optimization process is undergoing active development and may change.
- Our goal with these changes is to improve the performance of optimization and
- reduce the complexity.
-
- That being said, it is essential today to run optimize if you want the best
- performance.  It should be stable and safe to use in production, but it our
- hope that the API may be simplified (or not even need to be called) in the
- future.
-
- The frequency an application shoudl call optimize is based on the frequency of
+ The frequency an application should call optimize is based on the frequency of
 data modifications.  If data is frequently added, deleted, or updated then
 optimize should be run frequently.  A good rule of thumb is to run optimize if
 you have added or modified 100,000 or more records or run more than 20 data
--- a/docs/src/js/interfaces/OptimizeOptions.md
+++ b/docs/src/js/interfaces/OptimizeOptions.md
@@ -37,3 +37,12 @@ tbl.optimize({cleanupOlderThan: new Date()});
 ```ts
 deleteUnverified: boolean;
 ```
+
+Because they may be part of an in-progress transaction, files newer than
+7 days old are not deleted by default. If you are sure that there are no
+in-progress transactions, then you can set this to true to delete all
+files older than `cleanupOlderThan`.
+
+**WARNING**: This should only be set to true if you can guarantee that
+no other process is currently working on this dataset. Otherwise the
+dataset could be put into a corrupted state.
--- a/java/lancedb-core/pom.xml
+++ b/java/lancedb-core/pom.xml
@@ -8,7 +8,7 @@
    <parent>
      <groupId>com.lancedb</groupId>
      <artifactId>lancedb-parent</artifactId>
-      <version>0.27.0-beta.5</version>
+      <version>0.27.1-final.0</version>
      <relativePath>../pom.xml</relativePath>
    </parent>

--- a/java/pom.xml
+++ b/java/pom.xml
@@ -6,7 +6,7 @@

    <groupId>com.lancedb</groupId>
    <artifactId>lancedb-parent</artifactId>
-    <version>0.27.0-beta.5</version>
+    <version>0.27.1-final.0</version>
    <packaging>pom</packaging>
    <name>${project.artifactId}</name>
    <description>LanceDB Java SDK Parent POM</description>
@@ -28,7 +28,7 @@
    <properties>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <arrow.version>15.0.0</arrow.version>
-        <lance-core.version>3.1.0-beta.2</lance-core.version>
+        <lance-core.version>3.0.1</lance-core.version>
        <spotless.skip>false</spotless.skip>
        <spotless.version>2.30.0</spotless.version>
        <spotless.java.googlejavaformat.version>1.7</spotless.java.googlejavaformat.version>
--- a/nodejs/Cargo.toml
+++ b/nodejs/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.27.0-beta.5"
+version = "0.27.1"
 license.workspace = true
 description.workspace = true
 repository.workspace = true
--- a/nodejs/test/table.test.ts
+++ b/nodejs/test/table.test.ts
@@ -1259,6 +1259,98 @@ describe("schema evolution", function () {
    expect(await table.schema()).toEqual(expectedSchema);
  });

+  it("can add columns with schema for explicit data types", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Define schema for new columns with explicit data types
+    // Note: All columns must be nullable when using addColumns with Schema
+    // because they are initially populated with null values
+    const newColumnsSchema = new Schema([
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+      new Field("rating", new Int32(), true),
+    ]);
+
+    const result = await table.addColumns(newColumnsSchema);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+      new Field("rating", new Int32(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+
+    // Verify that new columns are populated with null values
+    const results = await table.query().toArray();
+    expect(results).toHaveLength(1);
+    expect(results[0].price).toBeNull();
+    expect(results[0].category).toBeNull();
+    expect(results[0].rating).toBeNull();
+  });
+
+  it("can add a single column using Field", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Add a single field
+    const priceField = new Field("price", new Float64(), true);
+    const result = await table.addColumns(priceField);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+  });
+
+  it("can add multiple columns using array of Fields", async function () {
+    const con = await connect(tmpDir.name);
+    const table = await con.createTable("vectors", [
+      { id: 1n, vector: [0.1, 0.2] },
+    ]);
+
+    // Add multiple fields as array
+    const fields = [
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+    ];
+    const result = await table.addColumns(fields);
+    expect(result).toHaveProperty("version");
+    expect(result.version).toBe(2);
+
+    const expectedSchema = new Schema([
+      new Field("id", new Int64(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float32(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+      new Field("category", new Utf8(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema);
+  });
+
  it("can alter the columns in the schema", async function () {
    const con = await connect(tmpDir.name);
    const schema = new Schema([
--- a/nodejs/lancedb/table.ts
+++ b/nodejs/lancedb/table.ts
@@ -5,12 +5,15 @@ import {
  Table as ArrowTable,
  Data,
  DataType,
+  Field,
  IntoVector,
  MultiVector,
  Schema,
  dataTypeToJson,
  fromDataToBuffer,
+  fromTableToBuffer,
  isMultiVector,
+  makeEmptyTable,
  tableFromIPC,
 } from "./arrow";

@@ -84,6 +87,16 @@ export interface OptimizeOptions {
   * tbl.optimize({cleanupOlderThan: new Date()});
   */
  cleanupOlderThan: Date;
+  /**
+   * Because they may be part of an in-progress transaction, files newer than
+   * 7 days old are not deleted by default. If you are sure that there are no
+   * in-progress transactions, then you can set this to true to delete all
+   * files older than `cleanupOlderThan`.
+   *
+   * **WARNING**: This should only be set to true if you can guarantee that
+   * no other process is currently working on this dataset. Otherwise the
+   * dataset could be put into a corrupted state.
+   */
  deleteUnverified: boolean;
 }

@@ -381,15 +394,16 @@ export abstract class Table {
  abstract vectorSearch(vector: IntoVector | MultiVector): VectorQuery;
  /**
   * Add new columns with defined values.
-   * @param {AddColumnsSql[]} newColumnTransforms pairs of column names and
-   * 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.
+   * @param {AddColumnsSql[] | Field | Field[] | Schema} newColumnTransforms Either:
+   *   - An array of objects with column names and SQL expressions to calculate values
+   *   - A single Arrow Field defining one column with its data type (column will be initialized with null values)
+   *   - An array of Arrow Fields defining columns with their data types (columns will be initialized with null values)
+   *   - An Arrow Schema defining columns with their data types (columns will be initialized with null values)
   * @returns {Promise<AddColumnsResult>} A promise that resolves to an object
   * containing the new version number of the table after adding the columns.
   */
  abstract addColumns(
-    newColumnTransforms: AddColumnsSql[],
+    newColumnTransforms: AddColumnsSql[] | Field | Field[] | Schema,
  ): Promise<AddColumnsResult>;

  /**
@@ -501,19 +515,7 @@ export abstract class Table {
   *  - Index: Optimizes the indices, adding new data to existing indices
   *
   *
-   *  Experimental API
-   *  ----------------
-   *
-   *  The optimization process is undergoing active development and may change.
-   *  Our goal with these changes is to improve the performance of optimization and
-   *  reduce the complexity.
-   *
-   *  That being said, it is essential today to run optimize if you want the best
-   *  performance.  It should be stable and safe to use in production, but it our
-   *  hope that the API may be simplified (or not even need to be called) in the
-   *  future.
-   *
-   *  The frequency an application shoudl call optimize is based on the frequency of
+   *  The frequency an application should call optimize is based on the frequency of
   *  data modifications.  If data is frequently added, deleted, or updated then
   *  optimize should be run frequently.  A good rule of thumb is to run optimize if
   *  you have added or modified 100,000 or more records or run more than 20 data
@@ -806,9 +808,40 @@ export class LocalTable extends Table {
  // TODO: Support BatchUDF

  async addColumns(
-    newColumnTransforms: AddColumnsSql[],
+    newColumnTransforms: AddColumnsSql[] | Field | Field[] | Schema,
  ): Promise<AddColumnsResult> {
-    return await this.inner.addColumns(newColumnTransforms);
+    // Handle single Field -> convert to array of Fields
+    if (newColumnTransforms instanceof Field) {
+      newColumnTransforms = [newColumnTransforms];
+    }
+
+    // Handle array of Fields -> convert to Schema
+    if (
+      Array.isArray(newColumnTransforms) &&
+      newColumnTransforms.length > 0 &&
+      newColumnTransforms[0] instanceof Field
+    ) {
+      const fields = newColumnTransforms as Field[];
+      newColumnTransforms = new Schema(fields);
+    }
+
+    // Handle Schema -> use schema-based approach
+    if (newColumnTransforms instanceof Schema) {
+      const schema = newColumnTransforms;
+      // Convert schema to buffer using Arrow IPC format
+      const emptyTable = makeEmptyTable(schema);
+      const schemaBuf = await fromTableToBuffer(emptyTable);
+      return await this.inner.addColumnsWithSchema(schemaBuf);
+    }
+
+    // Handle SQL expressions (existing functionality)
+    if (Array.isArray(newColumnTransforms)) {
+      return await this.inner.addColumns(
+        newColumnTransforms as AddColumnsSql[],
+      );
+    }
+
+    throw new Error("Invalid input type for addColumns");
  }

  async alterColumns(
--- a/nodejs/npm/darwin-arm64/package.json
+++ b/nodejs/npm/darwin-arm64/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",
--- a/nodejs/npm/linux-arm64-gnu/package.json
+++ b/nodejs/npm/linux-arm64-gnu/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",
--- a/nodejs/npm/linux-arm64-musl/package.json
+++ b/nodejs/npm/linux-arm64-musl/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",
--- a/nodejs/npm/linux-x64-gnu/package.json
+++ b/nodejs/npm/linux-x64-gnu/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",
--- a/nodejs/npm/linux-x64-musl/package.json
+++ b/nodejs/npm/linux-x64-musl/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",
--- a/nodejs/npm/win32-arm64-msvc/package.json
+++ b/nodejs/npm/win32-arm64-msvc/package.json
@@ -1,6 +1,6 @@
 {
  "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.27.0-beta.5",
+  "version": "0.27.1",
  "os": [
    "win32"
  ],
--- a/nodejs/npm/win32-x64-msvc/package.json
+++ b/nodejs/npm/win32-x64-msvc/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.27.0-beta.5",
+	"version": "0.27.1",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",
--- a/nodejs/package-lock.json
+++ b/nodejs/package-lock.json
@@ -1,12 +1,12 @@
 {
  "name": "@lancedb/lancedb",
-  "version": "0.27.0-beta.5",
+  "version": "0.27.1",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "@lancedb/lancedb",
-      "version": "0.27.0-beta.5",
+      "version": "0.27.1",
      "cpu": [
        "x64",
        "arm64"
--- a/nodejs/package.json
+++ b/nodejs/package.json
@@ -11,7 +11,7 @@
    "ann"
  ],
  "private": false,
-  "version": "0.27.0-beta.5",
+  "version": "0.27.1",
  "main": "dist/index.js",
  "exports": {
    ".": "./dist/index.js",
--- a/nodejs/src/table.rs
+++ b/nodejs/src/table.rs
@@ -3,7 +3,7 @@

 use std::collections::HashMap;

-use lancedb::ipc::ipc_file_to_batches;
+use lancedb::ipc::{ipc_file_to_batches, ipc_file_to_schema};
 use lancedb::table::{
    AddDataMode, ColumnAlteration as LanceColumnAlteration, Duration, NewColumnTransform,
    OptimizeAction, OptimizeOptions, Table as LanceDbTable,
@@ -279,6 +279,23 @@ impl Table {
        Ok(res.into())
    }

+    #[napi(catch_unwind)]
+    pub async fn add_columns_with_schema(
+        &self,
+        schema_buf: Buffer,
+    ) -> napi::Result<AddColumnsResult> {
+        let schema = ipc_file_to_schema(schema_buf.to_vec())
+            .map_err(|e| napi::Error::from_reason(format!("Failed to read IPC schema: {}", e)))?;
+
+        let transforms = NewColumnTransform::AllNulls(schema);
+        let res = self
+            .inner_ref()?
+            .add_columns(transforms, None)
+            .await
+            .default_error()?;
+        Ok(res.into())
+    }
+
    #[napi(catch_unwind)]
    pub async fn alter_columns(
        &self,
--- a/python/.bumpversion.toml
+++ b/python/.bumpversion.toml
@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.30.0-beta.5"
+current_version = "0.30.2-beta.0"
 parse = """(?x)
    (?P<major>0|[1-9]\\d*)\\.
    (?P<minor>0|[1-9]\\d*)\\.
--- a/python/Cargo.toml
+++ b/python/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.30.0-beta.5"
+version = "0.30.2-beta.0"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true
@@ -23,6 +23,7 @@ lance-namespace.workspace = true
 lance-namespace-impls.workspace = true
 lance-io.workspace = true
 env_logger.workspace = true
+log.workspace = true
 pyo3 = { version = "0.26", features = ["extension-module", "abi3-py39"] }
 pyo3-async-runtimes = { version = "0.26", features = [
    "attributes",
--- a/python/python/lancedb/_lancedb.pyi
+++ b/python/python/lancedb/_lancedb.pyi
@@ -135,7 +135,10 @@ class Table:
    def close(self) -> None: ...
    async def schema(self) -> pa.Schema: ...
    async def add(
-        self, data: pa.RecordBatchReader, mode: Literal["append", "overwrite"]
+        self,
+        data: pa.RecordBatchReader,
+        mode: Literal["append", "overwrite"],
+        progress: Optional[Any] = None,
    ) -> AddResult: ...
    async def update(
        self, updates: Dict[str, str], where: Optional[str]
--- a/python/python/lancedb/query.py
+++ b/python/python/lancedb/query.py
@@ -2205,8 +2205,8 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
            self._vector_query.select(self._columns)
            self._fts_query.select(self._columns)
        if self._where:
-            self._vector_query.where(self._where, self._postfilter)
-            self._fts_query.where(self._where, self._postfilter)
+            self._vector_query.where(self._where, not self._postfilter)
+            self._fts_query.where(self._where, not self._postfilter)
        if self._with_row_id:
            self._vector_query.with_row_id(True)
            self._fts_query.with_row_id(True)
--- a/python/python/lancedb/remote/table.py
+++ b/python/python/lancedb/remote/table.py
@@ -4,7 +4,7 @@
 from datetime import timedelta
 import logging
 from functools import cached_property
-from typing import Dict, Iterable, List, Optional, Union, Literal
+from typing import Any, Callable, Dict, Iterable, List, Optional, Union, Literal
 import warnings

 from lancedb._lancedb import (
@@ -35,6 +35,7 @@ import pyarrow as pa
 from lancedb.common import DATA, VEC, VECTOR_COLUMN_NAME
 from lancedb.merge import LanceMergeInsertBuilder
 from lancedb.embeddings import EmbeddingFunctionRegistry
+from lancedb.table import _normalize_progress

 from ..query import LanceVectorQueryBuilder, LanceQueryBuilder, LanceTakeQueryBuilder
 from ..table import AsyncTable, IndexStatistics, Query, Table, Tags
@@ -308,6 +309,7 @@ class RemoteTable(Table):
        mode: str = "append",
        on_bad_vectors: str = "error",
        fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
    ) -> AddResult:
        """Add more data to the [Table](Table). It has the same API signature as
        the OSS version.
@@ -330,17 +332,29 @@ class RemoteTable(Table):
            One of "error", "drop", "fill".
        fill_value: float, default 0.
            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.

        Returns
        -------
        AddResult
            An object containing the new version number of the table after adding data.
        """
-        return LOOP.run(
-            self._table.add(
-                data, mode=mode, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        progress, owns = _normalize_progress(progress)
+        try:
+            return LOOP.run(
+                self._table.add(
+                    data,
+                    mode=mode,
+                    on_bad_vectors=on_bad_vectors,
+                    fill_value=fill_value,
+                    progress=progress,
+                )
            )
-        )
+        finally:
+            if owns:
+                progress.close()

    def search(
        self,
--- a/python/python/lancedb/table.py
+++ b/python/python/lancedb/table.py
@@ -14,6 +14,7 @@ from functools import cached_property
 from typing import (
    TYPE_CHECKING,
    Any,
+    Callable,
    Dict,
    Iterable,
    List,
@@ -556,6 +557,21 @@ def _table_uri(base: str, table_name: str) -> str:
    return join_uri(base, f"{table_name}.lance")


+def _normalize_progress(progress):
+    """Normalize a ``progress`` parameter for :meth:`Table.add`.
+
+    Returns ``(progress_obj, owns)`` where *owns* is True when we created a
+    tqdm bar that the caller must close.
+    """
+    if progress is True:
+        from tqdm.auto import tqdm
+
+        return tqdm(unit=" rows"), True
+    if progress is False or progress is None:
+        return None, False
+    return progress, False
+
+
 class Table(ABC):
    """
    A Table is a collection of Records in a LanceDB Database.
@@ -974,6 +990,7 @@ class Table(ABC):
        mode: AddMode = "append",
        on_bad_vectors: OnBadVectorsType = "error",
        fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
    ) -> AddResult:
        """Add more data to the [Table](Table).

@@ -995,6 +1012,29 @@ class Table(ABC):
            One of "error", "drop", "fill".
        fill_value: float, default 0.
            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            Progress reporting during the add operation. Can be:
+
+            - ``True`` to automatically create and display a tqdm progress
+              bar (requires ``tqdm`` to be installed)::
+
+                table.add(data, progress=True)
+
+            - A **callable** that receives a dict with keys ``output_rows``,
+              ``output_bytes``, ``total_rows``, ``elapsed_seconds``,
+              ``active_tasks``, ``total_tasks``, and ``done``::
+
+                def on_progress(p):
+                    print(f"{p['output_rows']}/{p['total_rows']} rows, "
+                          f"{p['active_tasks']}/{p['total_tasks']} workers")
+                table.add(data, progress=on_progress)
+
+            - A **tqdm-compatible** progress bar whose ``total`` and
+              ``update()`` will be called automatically. The postfix shows
+              write throughput (MB/s) and active worker count::
+
+                with tqdm() as pbar:
+                    table.add(data, progress=pbar)

        Returns
        -------
@@ -1506,22 +1546,17 @@ class Table(ABC):
            in-progress operation (e.g. appending new data) and these files will not
            be deleted unless they are at least 7 days old. If delete_unverified is True
            then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
        retrain: bool, default False
            This parameter is no longer used and is deprecated.

-        Experimental API
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
+        The frequency an application should call optimize is based on the frequency of
        data modifications.  If data is frequently added, deleted, or updated then
        optimize should be run frequently.  A good rule of thumb is to run optimize if
        you have added or modified 100,000 or more records or run more than 20 data
@@ -2497,6 +2532,7 @@ class LanceTable(Table):
        mode: AddMode = "append",
        on_bad_vectors: OnBadVectorsType = "error",
        fill_value: float = 0.0,
+        progress: Optional[Union[bool, Callable, Any]] = None,
    ) -> AddResult:
        """Add data to the table.
        If vector columns are missing and the table
@@ -2515,17 +2551,29 @@ class LanceTable(Table):
            One of "error", "drop", "fill", "null".
        fill_value: float, default 0.
            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: bool, callable, or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.

        Returns
        -------
        int
            The number of vectors in the table.
        """
-        return LOOP.run(
-            self._table.add(
-                data, mode=mode, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        progress, owns = _normalize_progress(progress)
+        try:
+            return LOOP.run(
+                self._table.add(
+                    data,
+                    mode=mode,
+                    on_bad_vectors=on_bad_vectors,
+                    fill_value=fill_value,
+                    progress=progress,
+                )
            )
-        )
+        finally:
+            if owns:
+                progress.close()

    def merge(
        self,
@@ -3047,22 +3095,17 @@ class LanceTable(Table):
            in-progress operation (e.g. appending new data) and these files will not
            be deleted unless they are at least 7 days old. If delete_unverified is True
            then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
        retrain: bool, default False
            This parameter is no longer used and is deprecated.

-        Experimental API
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
+        The frequency an application should call optimize is based on the frequency of
        data modifications.  If data is frequently added, deleted, or updated then
        optimize should be run frequently.  A good rule of thumb is to run optimize if
        you have added or modified 100,000 or more records or run more than 20 data
@@ -3779,6 +3822,7 @@ class AsyncTable:
        mode: Optional[Literal["append", "overwrite"]] = "append",
        on_bad_vectors: Optional[OnBadVectorsType] = None,
        fill_value: Optional[float] = None,
+        progress: Optional[Union[bool, Callable, Any]] = None,
    ) -> AddResult:
        """Add more data to the [Table](Table).

@@ -3800,6 +3844,9 @@ class AsyncTable:
            One of "error", "drop", "fill", "null".
        fill_value: float, default 0.
            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        progress: callable or tqdm-like, optional
+            A callback or tqdm-compatible progress bar. See
+            :meth:`Table.add` for details.

        """
        schema = await self.schema()
@@ -3823,8 +3870,9 @@ class AsyncTable:
            )
        _register_optional_converters()
        data = to_scannable(data)
+        progress, owns = _normalize_progress(progress)
        try:
-            return await self._inner.add(data, mode or "append")
+            return await self._inner.add(data, mode or "append", progress=progress)
        except RuntimeError as e:
            if "Cast error" in str(e):
                raise ValueError(e)
@@ -3832,6 +3880,9 @@ class AsyncTable:
                raise ValueError(e)
            else:
                raise
+        finally:
+            if owns:
+                progress.close()

    def merge_insert(self, on: Union[str, Iterable[str]]) -> LanceMergeInsertBuilder:
        """
@@ -4630,22 +4681,17 @@ class AsyncTable:
            in-progress operation (e.g. appending new data) and these files will not
            be deleted unless they are at least 7 days old. If delete_unverified is True
            then these files will be deleted regardless of their age.
+
+            .. warning::
+
+                This should only be set to True if you can guarantee that no other
+                process is currently working on this dataset. Otherwise the dataset
+                could be put into a corrupted state.
+
        retrain: bool, default False
            This parameter is no longer used and is deprecated.

-        Experimental API
-        ----------------
-
-        The optimization process is undergoing active development and may change.
-        Our goal with these changes is to improve the performance of optimization and
-        reduce the complexity.
-
-        That being said, it is essential today to run optimize if you want the best
-        performance.  It should be stable and safe to use in production, but it our
-        hope that the API may be simplified (or not even need to be called) in the
-        future.
-
-        The frequency an application shoudl call optimize is based on the frequency of
+        The frequency an application should call optimize is based on the frequency of
        data modifications.  If data is frequently added, deleted, or updated then
        optimize should be run frequently.  A good rule of thumb is to run optimize if
        you have added or modified 100,000 or more records or run more than 20 data
@@ -4766,7 +4812,16 @@ class IndexStatistics:
    num_indexed_rows: int
    num_unindexed_rows: int
    index_type: Literal[
-        "IVF_PQ", "IVF_HNSW_PQ", "IVF_HNSW_SQ", "FTS", "BTREE", "BITMAP", "LABEL_LIST"
+        "IVF_FLAT",
+        "IVF_SQ",
+        "IVF_PQ",
+        "IVF_RQ",
+        "IVF_HNSW_SQ",
+        "IVF_HNSW_PQ",
+        "FTS",
+        "BTREE",
+        "BITMAP",
+        "LABEL_LIST",
    ]
    distance_type: Optional[Literal["l2", "cosine", "dot"]] = None
    num_indices: Optional[int] = None
--- a/python/python/tests/test_hybrid_query.py
+++ b/python/python/tests/test_hybrid_query.py
@@ -177,6 +177,60 @@ async def test_analyze_plan(table: AsyncTable):
    assert "metrics=" in res


+@pytest.fixture
+def table_with_id(tmpdir_factory) -> Table:
+    tmp_path = str(tmpdir_factory.mktemp("data"))
+    db = lancedb.connect(tmp_path)
+    data = pa.table(
+        {
+            "id": pa.array([1, 2, 3, 4], type=pa.int64()),
+            "text": pa.array(["a", "b", "cat", "dog"]),
+            "vector": pa.array(
+                [[0.1, 0.1], [2, 2], [-0.1, -0.1], [0.5, -0.5]],
+                type=pa.list_(pa.float32(), list_size=2),
+            ),
+        }
+    )
+    table = db.create_table("test_with_id", data)
+    table.create_fts_index("text", with_position=False, use_tantivy=False)
+    return table
+
+
+def test_hybrid_prefilter_explain_plan(table_with_id: Table):
+    """
+    Verify that the prefilter logic is not inverted in LanceHybridQueryBuilder.
+    """
+    plan_prefilter = (
+        table_with_id.search(query_type="hybrid")
+        .vector([0.0, 0.0])
+        .text("dog")
+        .where("id = 1", prefilter=True)
+        .limit(2)
+        .explain_plan(verbose=True)
+    )
+
+    plan_postfilter = (
+        table_with_id.search(query_type="hybrid")
+        .vector([0.0, 0.0])
+        .text("dog")
+        .where("id = 1", prefilter=False)
+        .limit(2)
+        .explain_plan(verbose=True)
+    )
+
+    # prefilter=True: filter is pushed into the LanceRead scan.
+    # The FTS sub-plan exposes this as "full_filter=id = Int64(1)" inside LanceRead.
+    assert "full_filter=id = Int64(1)" in plan_prefilter, (
+        f"Should push the filter into the scan.\nPlan:\n{plan_prefilter}"
+    )
+
+    # prefilter=False: filter is applied as a separate FilterExec after the search.
+    # The filter must NOT be embedded in the scan.
+    assert "full_filter=id = Int64(1)" not in plan_postfilter, (
+        f"Should NOT push the filter into the scan.\nPlan:\n{plan_postfilter}"
+    )
+
+
 def test_normalize_scores():
    cases = [
        (pa.array([0.1, 0.4]), pa.array([0.0, 1.0])),
--- a/python/python/tests/test_index.py
+++ b/python/python/tests/test_index.py
@@ -3,6 +3,7 @@

 from datetime import timedelta
 import random
+from typing import get_args, get_type_hints

 import pyarrow as pa
 import pytest
@@ -22,6 +23,7 @@ from lancedb.index import (
    HnswSq,
    FTS,
 )
+from lancedb.table import IndexStatistics


@pytest_asyncio.fixture
@@ -283,3 +285,23 @@ async def test_create_index_with_binary_vectors(binary_table: AsyncTable):
    for v in range(256):
        res = await binary_table.query().nearest_to([v] * 128).to_arrow()
        assert res["id"][0].as_py() == v
+
+
+def test_index_statistics_index_type_lists_all_supported_values():
+    expected_index_types = {
+        "IVF_FLAT",
+        "IVF_SQ",
+        "IVF_PQ",
+        "IVF_RQ",
+        "IVF_HNSW_SQ",
+        "IVF_HNSW_PQ",
+        "FTS",
+        "BTREE",
+        "BITMAP",
+        "LABEL_LIST",
+    }
+
+    assert (
+        set(get_args(get_type_hints(IndexStatistics)["index_type"]))
+        == expected_index_types
+    )
--- a/python/python/tests/test_namespace_integration.py
+++ b/python/python/tests/test_namespace_integration.py
@@ -147,7 +147,12 @@ class TrackingNamespace(LanceNamespace):
        This simulates a credential rotation system where each call returns
        new credentials that expire after credential_expires_in_seconds.
        """
-        modified = copy.deepcopy(storage_options) if storage_options else {}
+        # Start from base storage options (endpoint, region, allow_http, etc.)
+        # because DirectoryNamespace returns None for storage_options from
+        # describe_table/declare_table when no credential vendor is configured.
+        modified = copy.deepcopy(self.base_storage_options)
+        if storage_options:
+            modified.update(storage_options)

        # Increment credentials to simulate rotation
        modified["aws_access_key_id"] = f"AKID_{count}"
--- a/python/python/tests/test_table.py
+++ b/python/python/tests/test_table.py
@@ -527,6 +527,102 @@ async def test_add_async(mem_db_async: AsyncConnection):
    assert await table.count_rows() == 3


+def test_add_progress_callback(mem_db: DBConnection):
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    updates = []
+    table.add([{"id": 3}, {"id": 4}], progress=lambda p: updates.append(dict(p)))
+
+    assert len(table) == 4
+    # The done callback always fires, so we should always get at least one.
+    assert len(updates) >= 1, "expected at least one progress callback"
+    for p in updates:
+        assert "output_rows" in p
+        assert "output_bytes" in p
+        assert "total_rows" in p
+        assert "elapsed_seconds" in p
+        assert "active_tasks" in p
+        assert "total_tasks" in p
+        assert "done" in p
+    # The last callback should have done=True.
+    assert updates[-1]["done"] is True
+
+
+def test_add_progress_tqdm_like(mem_db: DBConnection):
+    """Test that a tqdm-like object gets total set and update() called."""
+
+    class FakeBar:
+        def __init__(self):
+            self.total = None
+            self.n = 0
+            self.postfix = None
+
+        def update(self, n):
+            self.n += n
+
+        def set_postfix_str(self, s):
+            self.postfix = s
+
+        def refresh(self):
+            pass
+
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    bar = FakeBar()
+    table.add([{"id": 3}, {"id": 4}], progress=bar)
+
+    assert len(table) == 4
+    # Postfix should contain throughput and worker count
+    if bar.postfix is not None:
+        assert "MB/s" in bar.postfix
+        assert "workers" in bar.postfix
+
+
+def test_add_progress_bool(mem_db: DBConnection):
+    """Test that progress=True creates and closes a tqdm bar automatically."""
+    table = mem_db.create_table(
+        "test",
+        data=[{"id": 1}, {"id": 2}],
+    )
+
+    table.add([{"id": 3}, {"id": 4}], progress=True)
+    assert len(table) == 4
+
+    # progress=False should be the same as None
+    table.add([{"id": 5}], progress=False)
+    assert len(table) == 5
+
+
+@pytest.mark.asyncio
+async def test_add_progress_callback_async(mem_db_async: AsyncConnection):
+    """Progress callbacks work through the async path too."""
+    table = await mem_db_async.create_table("test", data=[{"id": 1}, {"id": 2}])
+
+    updates = []
+    await table.add([{"id": 3}, {"id": 4}], progress=lambda p: updates.append(dict(p)))
+
+    assert await table.count_rows() == 4
+    assert len(updates) >= 1
+    assert updates[-1]["done"] is True
+
+
+def test_add_progress_callback_error(mem_db: DBConnection):
+    """A failing callback must not prevent the write from succeeding."""
+    table = mem_db.create_table("test", data=[{"id": 1}, {"id": 2}])
+
+    def bad_callback(p):
+        raise RuntimeError("boom")
+
+    table.add([{"id": 3}, {"id": 4}], progress=bad_callback)
+    assert len(table) == 4
+
+
 def test_polars(mem_db: DBConnection):
    data = {
        "vector": [[3.1, 4.1], [5.9, 26.5]],
--- a/python/src/query.rs
+++ b/python/src/query.rs
@@ -316,6 +316,19 @@ impl<'py> IntoPyObject<'py> for PySelect {
            Select::All => Ok(py.None().into_bound(py).into_any()),
            Select::Columns(columns) => Ok(columns.into_pyobject(py)?.into_any()),
            Select::Dynamic(columns) => Ok(columns.into_pyobject(py)?.into_any()),
+            Select::Expr(pairs) => {
+                // Serialize DataFusion Expr -> SQL string so Python sees the same
+                // format as Select::Dynamic: a list of (name, sql_string) tuples.
+                let sql_pairs: PyResult<Vec<(String, String)>> = pairs
+                    .into_iter()
+                    .map(|(name, expr)| {
+                        lancedb::expr::expr_to_sql_string(&expr)
+                            .map(|sql| (name, sql))
+                            .map_err(|e| PyRuntimeError::new_err(e.to_string()))
+                    })
+                    .collect();
+                Ok(sql_pairs?.into_pyobject(py)?.into_any())
+            }
        }
    }
 }
--- a/python/src/table.rs
+++ b/python/src/table.rs
@@ -19,7 +19,7 @@ use lancedb::table::{
    Table as LanceDbTable,
 };
 use pyo3::{
-    Bound, FromPyObject, PyAny, PyRef, PyResult, Python,
+    Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
    exceptions::{PyKeyError, PyRuntimeError, PyValueError},
    pyclass, pymethods,
    types::{IntoPyDict, PyAnyMethods, PyDict, PyDictMethods},
@@ -299,10 +299,12 @@ impl Table {
        })
    }

+    #[pyo3(signature = (data, mode, progress=None))]
    pub fn add<'a>(
        self_: PyRef<'a, Self>,
        data: PyScannable,
        mode: String,
+        progress: Option<Py<PyAny>>,
    ) -> PyResult<Bound<'a, PyAny>> {
        let mut op = self_.inner_ref()?.add(data);
        if mode == "append" {
@@ -312,6 +314,81 @@ impl Table {
        } else {
            return Err(PyValueError::new_err(format!("Invalid mode: {}", mode)));
        }
+        if let Some(progress_obj) = progress {
+            let is_callable = Python::attach(|py| progress_obj.bind(py).is_callable());
+            if is_callable {
+                // Callback: call with a dict of progress info.
+                op = op.progress(move |p| {
+                    Python::attach(|py| {
+                        let dict = PyDict::new(py);
+                        if let Err(e) = dict
+                            .set_item("output_rows", p.output_rows())
+                            .and_then(|_| dict.set_item("output_bytes", p.output_bytes()))
+                            .and_then(|_| dict.set_item("total_rows", p.total_rows()))
+                            .and_then(|_| {
+                                dict.set_item("elapsed_seconds", p.elapsed().as_secs_f64())
+                            })
+                            .and_then(|_| dict.set_item("active_tasks", p.active_tasks()))
+                            .and_then(|_| dict.set_item("total_tasks", p.total_tasks()))
+                            .and_then(|_| dict.set_item("done", p.done()))
+                        {
+                            log::warn!("progress dict error: {e}");
+                            return;
+                        }
+                        if let Err(e) = progress_obj.call1(py, (dict,)) {
+                            log::warn!("progress callback error: {e}");
+                        }
+                    });
+                });
+            } else {
+                // tqdm-like: has update() method.
+                let mut last_rows: usize = 0;
+                let mut total_set = false;
+                op = op.progress(move |p| {
+                    let current = p.output_rows();
+                    let prev = last_rows;
+                    last_rows = current;
+                    Python::attach(|py| {
+                        if let Some(total) = p.total_rows()
+                            && !total_set
+                        {
+                            if let Err(e) = progress_obj.setattr(py, "total", total) {
+                                log::warn!("progress setattr error: {e}");
+                            }
+                            total_set = true;
+                        }
+                        let delta = current.saturating_sub(prev);
+                        if delta > 0 {
+                            if let Err(e) = progress_obj.call_method1(py, "update", (delta,)) {
+                                log::warn!("progress update error: {e}");
+                            }
+                            // Show throughput and active workers in tqdm postfix.
+                            let elapsed = p.elapsed().as_secs_f64();
+                            if elapsed > 0.0 {
+                                let mb_per_sec = p.output_bytes() as f64 / elapsed / 1_000_000.0;
+                                let postfix = format!(
+                                    "{:.1} MB/s | {}/{} workers",
+                                    mb_per_sec,
+                                    p.active_tasks(),
+                                    p.total_tasks()
+                                );
+                                if let Err(e) =
+                                    progress_obj.call_method1(py, "set_postfix_str", (postfix,))
+                                {
+                                    log::warn!("progress set_postfix_str error: {e}");
+                                }
+                            }
+                        }
+                        if p.done() {
+                            // Force a final refresh so the bar shows completion.
+                            if let Err(e) = progress_obj.call_method0(py, "refresh") {
+                                log::warn!("progress refresh error: {e}");
+                            }
+                        }
+                    });
+                });
+            }
+        }

        future_into_py(self_.py(), async move {
            let result = op.execute().await.infer_error()?;
--- a/rust/lancedb/Cargo.toml
+++ b/rust/lancedb/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.27.0-beta.5"
+version = "0.27.1"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true
--- a/rust/lancedb/src/connection.rs
+++ b/rust/lancedb/src/connection.rs
@@ -596,11 +596,8 @@ pub struct ConnectBuilder {
 }

 #[cfg(feature = "remote")]
-const ENV_VARS_TO_STORAGE_OPTS: [(&str, &str); 3] = [
-    ("AZURE_STORAGE_ACCOUNT_NAME", "azure_storage_account_name"),
-    ("AZURE_CLIENT_ID", "azure_client_id"),
-    ("AZURE_TENANT_ID", "azure_tenant_id"),
-];
+const ENV_VARS_TO_STORAGE_OPTS: [(&str, &str); 1] =
+    [("AZURE_STORAGE_ACCOUNT_NAME", "azure_storage_account_name")];

 impl ConnectBuilder {
    /// Create a new [`ConnectOptions`] with the given database URI.
--- a/rust/lancedb/src/dataloader/permutation/reader.rs
+++ b/rust/lancedb/src/dataloader/permutation/reader.rs
@@ -339,6 +339,12 @@ impl PermutationReader {
                }
                Ok(false)
            }
+            Select::Expr(columns) => {
+                // For Expr projections, we check if any alias is _rowid.
+                // We can't validate the expression itself (it may differ from _rowid)
+                // but we allow it through; the column will be included.
+                Ok(columns.iter().any(|(alias, _)| alias == ROW_ID))
+            }
        }
    }

--- a/rust/lancedb/src/query.rs
+++ b/rust/lancedb/src/query.rs
@@ -47,6 +47,25 @@ pub enum Select {
    ///
    /// See [`Query::select`] for more details and examples
    Dynamic(Vec<(String, String)>),
+    /// Advanced selection using type-safe DataFusion expressions
+    ///
+    /// Similar to [`Select::Dynamic`] but uses [`datafusion_expr::Expr`] instead of
+    /// raw SQL strings. Use [`crate::expr`] helpers to build expressions:
+    ///
+    /// ```
+    /// use lancedb::expr::{col, lit};
+    /// use lancedb::query::Select;
+    ///
+    /// // SELECT id, id * 2 AS id2 FROM ...
+    /// let selection = Select::expr_projection(&[
+    ///     ("id", col("id")),
+    ///     ("id2", col("id") * lit(2)),
+    /// ]);
+    /// ```
+    ///
+    /// Note: For remote/server-side queries the expressions are serialized to SQL strings
+    /// automatically (same as [`Select::Dynamic`]).
+    Expr(Vec<(String, datafusion_expr::Expr)>),
 }

 impl Select {
@@ -69,6 +88,29 @@ impl Select {
                .collect(),
        )
    }
+    /// Create a typed-expression projection.
+    ///
+    /// This is a convenience method for creating a [`Select::Expr`] variant from
+    /// a slice of `(name, expr)` pairs where each `expr` is a [`datafusion_expr::Expr`].
+    ///
+    /// # Example
+    /// ```
+    /// use lancedb::expr::{col, lit};
+    /// use lancedb::query::Select;
+    ///
+    /// let selection = Select::expr_projection(&[
+    ///     ("id", col("id")),
+    ///     ("id2", col("id") * lit(2)),
+    /// ]);
+    /// ```
+    pub fn expr_projection(columns: &[(impl AsRef<str>, datafusion_expr::Expr)]) -> Self {
+        Self::Expr(
+            columns
+                .iter()
+                .map(|(name, expr)| (name.as_ref().to_string(), expr.clone()))
+                .collect(),
+        )
+    }
 }

 /// A trait for converting a type to a query vector
@@ -1591,6 +1633,58 @@ mod tests {
        });
    }

+    #[tokio::test]
+    async fn test_select_with_expr_projection() {
+        // Mirrors test_select_with_transform but uses Select::Expr instead of Select::Dynamic
+        let tmp_dir = tempdir().unwrap();
+        let dataset_path = tmp_dir.path().join("test_expr.lance");
+        let uri = dataset_path.to_str().unwrap();
+
+        let batches = make_non_empty_batches();
+        let conn = connect(uri).execute().await.unwrap();
+        let table = conn
+            .create_table("my_table", batches)
+            .execute()
+            .await
+            .unwrap();
+
+        use crate::expr::{col, lit};
+        let query = table.query().limit(10).select(Select::expr_projection(&[
+            ("id2", col("id") * lit(2i32)),
+            ("id", col("id")),
+        ]));
+
+        let schema = query.output_schema().await.unwrap();
+        assert_eq!(
+            schema,
+            Arc::new(ArrowSchema::new(vec![
+                ArrowField::new("id2", DataType::Int32, true),
+                ArrowField::new("id", DataType::Int32, true),
+            ]))
+        );
+
+        let result = query.execute().await;
+        let mut batches = result
+            .expect("should have result")
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+        assert_eq!(batches.len(), 1);
+        let batch = batches.pop().unwrap();
+
+        // id and id2
+        assert_eq!(batch.num_columns(), 2);
+
+        let id: &Int32Array = batch.column_by_name("id").unwrap().as_primitive();
+        let id2: &Int32Array = batch.column_by_name("id2").unwrap().as_primitive();
+
+        id.iter().zip(id2.iter()).for_each(|(id, id2)| {
+            let id = id.unwrap();
+            let id2 = id2.unwrap();
+            assert_eq!(id * 2, id2);
+        });
+    }
+
    #[tokio::test]
    async fn test_execute_no_vector() {
        // TODO: Switch back to memory://foo after https://github.com/lancedb/lancedb/issues/1051
--- a/rust/lancedb/src/remote/client.rs
+++ b/rust/lancedb/src/remote/client.rs
@@ -443,23 +443,13 @@ impl<S: HttpSend> RestfulLanceDbClient<S> {
                })?,
            );
        }
-        // Map azure storage options to x-azure-* headers.
-        // The option key uses underscores (e.g. "azure_client_id") while the
-        // header uses hyphens (e.g. "x-azure-client-id").
-        let azure_opts: [(&str, &str); 3] = [
-            ("azure_storage_account_name", "x-azure-storage-account-name"),
-            ("azure_client_id", "x-azure-client-id"),
-            ("azure_tenant_id", "x-azure-tenant-id"),
-        ];
-        for (opt_key, header_name) in azure_opts {
-            if let Some(v) = options.0.get(opt_key) {
-                headers.insert(
-                    HeaderName::from_static(header_name),
-                    HeaderValue::from_str(v).map_err(|_| Error::InvalidInput {
-                        message: format!("non-ascii value '{}' for option '{}'", v, opt_key),
-                    })?,
-                );
-            }
+        if let Some(v) = options.0.get("azure_storage_account_name") {
+            headers.insert(
+                HeaderName::from_static("x-azure-storage-account-name"),
+                HeaderValue::from_str(v).map_err(|_| Error::InvalidInput {
+                    message: format!("non-ascii storage account name '{}' provided", db_name),
+                })?,
+            );
        }

        for (key, value) in &config.extra_headers {
@@ -1082,34 +1072,4 @@ mod tests {
            _ => panic!("Expected Runtime error"),
        }
    }
-
-    #[test]
-    fn test_default_headers_azure_opts() {
-        let mut opts = HashMap::new();
-        opts.insert(
-            "azure_storage_account_name".to_string(),
-            "myaccount".to_string(),
-        );
-        opts.insert("azure_client_id".to_string(), "my-client-id".to_string());
-        opts.insert("azure_tenant_id".to_string(), "my-tenant-id".to_string());
-        let remote_opts = RemoteOptions::new(opts);
-
-        let headers = RestfulLanceDbClient::<Sender>::default_headers(
-            "test-key",
-            "us-east-1",
-            "testdb",
-            false,
-            &remote_opts,
-            None,
-            &ClientConfig::default(),
-        )
-        .unwrap();
-
-        assert_eq!(
-            headers.get("x-azure-storage-account-name").unwrap(),
-            "myaccount"
-        );
-        assert_eq!(headers.get("x-azure-client-id").unwrap(), "my-client-id");
-        assert_eq!(headers.get("x-azure-tenant-id").unwrap(), "my-tenant-id");
-    }
 }
--- a/rust/lancedb/src/remote/db.rs
+++ b/rust/lancedb/src/remote/db.rs
@@ -72,6 +72,10 @@ impl ServerVersion {
    pub fn support_structural_fts(&self) -> bool {
        self.0 >= semver::Version::new(0, 3, 0)
    }
+
+    pub fn support_multipart_write(&self) -> bool {
+        self.0 >= semver::Version::new(0, 4, 0)
+    }
 }

 pub const OPT_REMOTE_PREFIX: &str = "remote_database_";
@@ -778,12 +782,7 @@ impl RemoteOptions {

 impl From<StorageOptions> for RemoteOptions {
    fn from(options: StorageOptions) -> Self {
-        let supported_opts = vec![
-            "account_name",
-            "azure_storage_account_name",
-            "azure_client_id",
-            "azure_tenant_id",
-        ];
+        let supported_opts = vec!["account_name", "azure_storage_account_name"];
        let mut filtered = HashMap::new();
        for opt in supported_opts {
            if let Some(v) = options.0.get(opt) {
--- a/rust/lancedb/src/remote/table.rs
+++ b/rust/lancedb/src/remote/table.rs
--- a/rust/lancedb/src/remote/table/insert.rs
+++ b/rust/lancedb/src/remote/table/insert.rs
@@ -11,10 +11,14 @@ use arrow_ipc::CompressionType;
 use datafusion_common::{DataFusionError, Result as DataFusionResult};
 use datafusion_execution::{SendableRecordBatchStream, TaskContext};
 use datafusion_physical_expr::EquivalenceProperties;
+use datafusion_physical_plan::metrics::{ExecutionPlanMetricsSet, MetricsSet};
 use datafusion_physical_plan::stream::RecordBatchStreamAdapter;
-use datafusion_physical_plan::{DisplayAs, DisplayFormatType, ExecutionPlan, PlanProperties};
+use datafusion_physical_plan::{
+    DisplayAs, DisplayFormatType, ExecutionPlan, ExecutionPlanProperties, PlanProperties,
+};
 use futures::StreamExt;
 use http::header::CONTENT_TYPE;
+use lance::io::exec::utils::InstrumentedRecordBatchStreamAdapter;

 use crate::Error;
 use crate::remote::ARROW_STREAM_CONTENT_TYPE;
@@ -22,13 +26,16 @@ use crate::remote::client::{HttpSend, RestfulLanceDbClient, Sender};
 use crate::remote::table::RemoteTable;
 use crate::table::AddResult;
 use crate::table::datafusion::insert::COUNT_SCHEMA;
+use crate::table::write_progress::WriteProgressTracker;

 /// ExecutionPlan for inserting data into a remote LanceDB table.
 ///
-/// This plan:
-/// 1. Requires single partition (no parallel remote inserts yet)
-/// 2. Streams data as Arrow IPC to `/v1/table/{id}/insert/` endpoint
-/// 3. Stores AddResult for retrieval after execution
+/// Streams data as Arrow IPC to `/v1/table/{id}/insert/` endpoint.
+///
+/// When `upload_id` is set, inserts are staged as part of a multipart write
+/// session and the plan supports multiple partitions for parallel uploads.
+/// Without `upload_id`, the plan requires a single partition and commits
+/// immediately.
 #[derive(Debug)]
 pub struct RemoteInsertExec<S: HttpSend = Sender> {
    table_name: String,
@@ -38,21 +45,69 @@ pub struct RemoteInsertExec<S: HttpSend = Sender> {
    overwrite: bool,
    properties: PlanProperties,
    add_result: Arc<Mutex<Option<AddResult>>>,
+    metrics: ExecutionPlanMetricsSet,
+    upload_id: Option<String>,
+    tracker: Option<Arc<WriteProgressTracker>>,
 }

 impl<S: HttpSend + 'static> RemoteInsertExec<S> {
-    /// Create a new RemoteInsertExec.
+    /// Create a new single-partition RemoteInsertExec.
    pub fn new(
        table_name: String,
        identifier: String,
        client: RestfulLanceDbClient<S>,
        input: Arc<dyn ExecutionPlan>,
        overwrite: bool,
+        tracker: Option<Arc<WriteProgressTracker>>,
    ) -> Self {
+        Self::new_inner(
+            table_name, identifier, client, input, overwrite, None, tracker,
+        )
+    }
+
+    /// Create a multi-partition RemoteInsertExec for use with multipart writes.
+    ///
+    /// Each partition's insert is staged under the given `upload_id` without
+    /// committing. The caller is responsible for calling the complete (or abort)
+    /// endpoint after all partitions finish.
+    pub fn new_multipart(
+        table_name: String,
+        identifier: String,
+        client: RestfulLanceDbClient<S>,
+        input: Arc<dyn ExecutionPlan>,
+        overwrite: bool,
+        upload_id: String,
+        tracker: Option<Arc<WriteProgressTracker>>,
+    ) -> Self {
+        Self::new_inner(
+            table_name,
+            identifier,
+            client,
+            input,
+            overwrite,
+            Some(upload_id),
+            tracker,
+        )
+    }
+
+    fn new_inner(
+        table_name: String,
+        identifier: String,
+        client: RestfulLanceDbClient<S>,
+        input: Arc<dyn ExecutionPlan>,
+        overwrite: bool,
+        upload_id: Option<String>,
+        tracker: Option<Arc<WriteProgressTracker>>,
+    ) -> Self {
+        let num_partitions = if upload_id.is_some() {
+            input.output_partitioning().partition_count()
+        } else {
+            1
+        };
        let schema = COUNT_SCHEMA.clone();
        let properties = PlanProperties::new(
            EquivalenceProperties::new(schema),
-            datafusion_physical_plan::Partitioning::UnknownPartitioning(1),
+            datafusion_physical_plan::Partitioning::UnknownPartitioning(num_partitions),
            datafusion_physical_plan::execution_plan::EmissionType::Final,
            datafusion_physical_plan::execution_plan::Boundedness::Bounded,
        );
@@ -65,6 +120,9 @@ impl<S: HttpSend + 'static> RemoteInsertExec<S> {
            overwrite,
            properties,
            add_result: Arc::new(Mutex::new(None)),
+            metrics: ExecutionPlanMetricsSet::new(),
+            upload_id,
+            tracker,
        }
    }

@@ -83,6 +141,7 @@ impl<S: HttpSend + 'static> RemoteInsertExec<S> {
    fn stream_as_http_body(
        data: SendableRecordBatchStream,
        error_tx: tokio::sync::oneshot::Sender<DataFusionError>,
+        tracker: Option<Arc<WriteProgressTracker>>,
    ) -> DataFusionResult<reqwest::Body> {
        let options = arrow_ipc::writer::IpcWriteOptions::default()
            .try_with_compression(Some(CompressionType::LZ4_FRAME))?;
@@ -94,37 +153,46 @@ impl<S: HttpSend + 'static> RemoteInsertExec<S> {

        let stream = futures::stream::try_unfold(
            (data, writer, Some(error_tx), false),
-            move |(mut data, mut writer, error_tx, finished)| async move {
-                if finished {
-                    return Ok(None);
-                }
-                match data.next().await {
-                    Some(Ok(batch)) => {
-                        writer
-                            .write(&batch)
-                            .map_err(|e| std::io::Error::other(e.to_string()))?;
-                        let buffer = std::mem::take(writer.get_mut());
-                        Ok(Some((buffer, (data, writer, error_tx, false))))
+            move |(mut data, mut writer, error_tx, finished)| {
+                let tracker = tracker.clone();
+                async move {
+                    if finished {
+                        return Ok(None);
                    }
-                    Some(Err(e)) => {
-                        // Send the original error through the channel before
-                        // returning a generic error to reqwest.
-                        if let Some(tx) = error_tx {
-                            let _ = tx.send(e);
+                    match data.next().await {
+                        Some(Ok(batch)) => {
+                            writer
+                                .write(&batch)
+                                .map_err(|e| std::io::Error::other(e.to_string()))?;
+                            let buffer = std::mem::take(writer.get_mut());
+                            if let Some(ref t) = tracker {
+                                t.record_bytes(buffer.len());
+                            }
+                            Ok(Some((buffer, (data, writer, error_tx, false))))
                        }
-                        Err(std::io::Error::other(
-                            "input stream error (see error channel)",
-                        ))
-                    }
-                    None => {
-                        writer
-                            .finish()
-                            .map_err(|e| std::io::Error::other(e.to_string()))?;
-                        let buffer = std::mem::take(writer.get_mut());
-                        if buffer.is_empty() {
-                            Ok(None)
-                        } else {
-                            Ok(Some((buffer, (data, writer, None, true))))
+                        Some(Err(e)) => {
+                            // Send the original error through the channel before
+                            // returning a generic error to reqwest.
+                            if let Some(tx) = error_tx {
+                                let _ = tx.send(e);
+                            }
+                            Err(std::io::Error::other(
+                                "input stream error (see error channel)",
+                            ))
+                        }
+                        None => {
+                            writer
+                                .finish()
+                                .map_err(|e| std::io::Error::other(e.to_string()))?;
+                            let buffer = std::mem::take(writer.get_mut());
+                            if buffer.is_empty() {
+                                Ok(None)
+                            } else {
+                                if let Some(ref t) = tracker {
+                                    t.record_bytes(buffer.len());
+                                }
+                                Ok(Some((buffer, (data, writer, None, true))))
+                            }
                        }
                    }
                }
@@ -174,8 +242,11 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
    }

    fn required_input_distribution(&self) -> Vec<datafusion_physical_plan::Distribution> {
-        // Until we have a separate commit endpoint, we need to do all inserts in a single partition
-        vec![datafusion_physical_plan::Distribution::SinglePartition]
+        if self.upload_id.is_some() {
+            vec![datafusion_physical_plan::Distribution::UnspecifiedDistribution]
+        } else {
+            vec![datafusion_physical_plan::Distribution::SinglePartition]
+        }
    }

    fn benefits_from_input_partitioning(&self) -> Vec<bool> {
@@ -191,12 +262,14 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
                "RemoteInsertExec requires exactly one child".to_string(),
            ));
        }
-        Ok(Arc::new(Self::new(
+        Ok(Arc::new(Self::new_inner(
            self.table_name.clone(),
            self.identifier.clone(),
            self.client.clone(),
            children[0].clone(),
            self.overwrite,
+            self.upload_id.clone(),
+            self.tracker.clone(),
        )))
    }

@@ -205,18 +278,29 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
        partition: usize,
        context: Arc<TaskContext>,
    ) -> DataFusionResult<SendableRecordBatchStream> {
-        if partition != 0 {
+        if self.upload_id.is_none() && partition != 0 {
            return Err(DataFusionError::Internal(
-                "RemoteInsertExec only supports single partition execution".to_string(),
+                "RemoteInsertExec only supports single partition execution without upload_id"
+                    .to_string(),
            ));
        }

-        let input_stream = self.input.execute(0, context)?;
+        let input_stream = self.input.execute(partition, context)?;
+        let input_schema = input_stream.schema();
+        let input_stream: SendableRecordBatchStream =
+            Box::pin(InstrumentedRecordBatchStreamAdapter::new(
+                input_schema,
+                input_stream,
+                partition,
+                &self.metrics,
+            ));
        let client = self.client.clone();
        let identifier = self.identifier.clone();
        let overwrite = self.overwrite;
        let add_result = self.add_result.clone();
        let table_name = self.table_name.clone();
+        let upload_id = self.upload_id.clone();
+        let tracker = self.tracker.clone();

        let stream = futures::stream::once(async move {
            let mut request = client
@@ -226,9 +310,12 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
            if overwrite {
                request = request.query(&[("mode", "overwrite")]);
            }
+            if let Some(ref uid) = upload_id {
+                request = request.query(&[("upload_id", uid.as_str())]);
+            }

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

            let result: DataFusionResult<(String, _)> = async {
@@ -262,32 +349,43 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {

            let (request_id, response) = result?;

-            let body_text = response.text().await.map_err(|e| {
-                DataFusionError::External(Box::new(Error::Http {
-                    source: Box::new(e),
-                    request_id: request_id.clone(),
-                    status_code: None,
-                }))
-            })?;
-
-            let parsed_result = if body_text.trim().is_empty() {
-                // Backward compatible with old servers
-                AddResult { version: 0 }
-            } else {
-                serde_json::from_str(&body_text).map_err(|e| {
+            // For multipart writes, the staging response is not the final
+            // version. Only parse AddResult for non-multipart inserts.
+            if upload_id.is_none() {
+                let body_text = response.text().await.map_err(|e| {
                    DataFusionError::External(Box::new(Error::Http {
-                        source: format!("Failed to parse add response: {}", e).into(),
+                        source: Box::new(e),
                        request_id: request_id.clone(),
                        status_code: None,
                    }))
-                })?
-            };
+                })?;
+
+                let parsed_result = if body_text.trim().is_empty() {
+                    // Backward compatible with old servers
+                    AddResult { version: 0 }
+                } else {
+                    serde_json::from_str(&body_text).map_err(|e| {
+                        DataFusionError::External(Box::new(Error::Http {
+                            source: format!("Failed to parse add response: {}", e).into(),
+                            request_id: request_id.clone(),
+                            status_code: None,
+                        }))
+                    })?
+                };

-            {
                let mut res_lock = add_result.lock().map_err(|_| {
                    DataFusionError::Execution("Failed to acquire lock for add_result".to_string())
                })?;
                *res_lock = Some(parsed_result);
+            } else {
+                // We don't use the body in this case, but we should still consume it.
+                let _ = response.bytes().await.map_err(|e| {
+                    DataFusionError::External(Box::new(Error::Http {
+                        source: Box::new(e),
+                        request_id: request_id.clone(),
+                        status_code: None,
+                    }))
+                })?;
            }

            // Return a single batch with count 0 (actual count is tracked in add_result)
@@ -301,6 +399,10 @@ impl<S: HttpSend + 'static> ExecutionPlan for RemoteInsertExec<S> {
            stream,
        )))
    }
+
+    fn metrics(&self) -> Option<MetricsSet> {
+        Some(self.metrics.clone_inner())
+    }
 }

 #[cfg(test)]
--- a/rust/lancedb/src/table.rs
+++ b/rust/lancedb/src/table.rs
@@ -74,7 +74,10 @@ pub mod optimize;
 pub mod query;
 pub mod schema_evolution;
 pub mod update;
+pub mod write_progress;
 use crate::index::waiter::wait_for_index;
+#[cfg(feature = "remote")]
+pub(crate) use add_data::PreprocessingOutput;
 pub use add_data::{AddDataBuilder, AddDataMode, AddResult, NaNVectorBehavior};
 pub use chrono::Duration;
 pub use delete::DeleteResult;
@@ -440,6 +443,34 @@ mod test_utils {
                embedding_registry: Arc::new(MemoryRegistry::new()),
            }
        }
+
+        pub fn new_with_handler_version_and_config<T>(
+            name: impl Into<String>,
+            version: semver::Version,
+            handler: impl Fn(reqwest::Request) -> http::Response<T> + Clone + Send + Sync + 'static,
+            config: crate::remote::ClientConfig,
+        ) -> Self
+        where
+            T: Into<reqwest::Body>,
+        {
+            let inner = Arc::new(
+                crate::remote::table::RemoteTable::new_mock_with_version_and_config(
+                    name.into(),
+                    handler.clone(),
+                    Some(version),
+                    config.clone(),
+                ),
+            );
+            let database = Arc::new(crate::remote::db::RemoteDatabase::new_mock_with_config(
+                handler, config,
+            ));
+            Self {
+                inner,
+                database: Some(database),
+                // Registry is unused.
+                embedding_registry: Arc::new(MemoryRegistry::new()),
+            }
+        }
    }
 }

@@ -951,17 +982,7 @@ impl Table {
    ///  * Prune: Removes old versions of the dataset
    ///  * Index: Optimizes the indices, adding new data to existing indices
    ///
-    /// <section class="warning">Experimental API</section>
-    ///
-    /// The optimization process is undergoing active development and may change.
-    /// Our goal with these changes is to improve the performance of optimization and
-    /// reduce the complexity.
-    ///
-    /// That being said, it is essential today to run optimize if you want the best
-    /// performance.  It should be stable and safe to use in production, but it our
-    /// hope that the API may be simplified (or not even need to be called) in the future.
-    ///
-    /// The frequency an application shoudl call optimize is based on the frequency of
+    /// The frequency an application should call optimize is based on the frequency of
    /// data modifications.  If data is frequently added, deleted, or updated then
    /// optimize should be run frequently.  A good rule of thumb is to run optimize if
    /// you have added or modified 100,000 or more records or run more than 20 data
@@ -2208,21 +2229,26 @@ impl BaseTable for NativeTable {

        let table_schema = Schema::from(&ds.schema().clone());

-        // Peek at the first batch to estimate a good partition count for
-        // write parallelism.
-        let mut peeked = PeekedScannable::new(add.data);
-        let num_partitions = if let Some(first_batch) = peeked.peek().await {
-            let max_partitions = lance_core::utils::tokio::get_num_compute_intensive_cpus();
-            estimate_write_partitions(
-                first_batch.get_array_memory_size(),
-                first_batch.num_rows(),
-                peeked.num_rows(),
-                max_partitions,
-            )
+        let num_partitions = if let Some(parallelism) = add.write_parallelism {
+            parallelism
        } else {
-            1
+            // Peek at the first batch to estimate a good partition count for
+            // write parallelism.
+            let mut peeked = PeekedScannable::new(add.data);
+            let n = if let Some(first_batch) = peeked.peek().await {
+                let max_partitions = lance_core::utils::tokio::get_num_compute_intensive_cpus();
+                estimate_write_partitions(
+                    first_batch.get_array_memory_size(),
+                    first_batch.num_rows(),
+                    peeked.num_rows(),
+                    max_partitions,
+                )
+            } else {
+                1
+            };
+            add.data = Box::new(peeked);
+            n
        };
-        add.data = Box::new(peeked);

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

@@ -2251,13 +2277,21 @@ impl BaseTable for NativeTable {

        let insert_exec = Arc::new(InsertExec::new(ds_wrapper.clone(), ds, plan, lance_params));

+        let tracker_for_tasks = output.tracker.clone();
+        if let Some(ref t) = tracker_for_tasks {
+            t.set_total_tasks(num_partitions);
+        }
+        let _finish = write_progress::FinishOnDrop(output.tracker);
+
        // Execute all partitions in parallel.
        let task_ctx = Arc::new(TaskContext::default());
        let handles = FuturesUnordered::new();
        for partition in 0..num_partitions {
            let exec = insert_exec.clone();
            let ctx = task_ctx.clone();
+            let tracker = tracker_for_tasks.clone();
            handles.push(tokio::spawn(async move {
+                let _guard = tracker.as_ref().map(|t| t.track_task());
                let mut stream = exec
                    .execute(partition, ctx)
                    .map_err(|e| -> Error { e.into() })?;
--- a/rust/lancedb/src/table/add_data.rs
+++ b/rust/lancedb/src/table/add_data.rs
@@ -13,6 +13,9 @@ use crate::embeddings::EmbeddingRegistry;
 use crate::table::datafusion::cast::cast_to_table_schema;
 use crate::table::datafusion::reject_nan::reject_nan_vectors;
 use crate::table::datafusion::scannable_exec::ScannableExec;
+use crate::table::write_progress::ProgressCallback;
+use crate::table::write_progress::WriteProgress;
+use crate::table::write_progress::WriteProgressTracker;
 use crate::{Error, Result};

 use super::{BaseTable, TableDefinition, WriteOptions};
@@ -52,6 +55,8 @@ pub struct AddDataBuilder {
    pub(crate) write_options: WriteOptions,
    pub(crate) on_nan_vectors: NaNVectorBehavior,
    pub(crate) embedding_registry: Option<Arc<dyn EmbeddingRegistry>>,
+    pub(crate) progress_callback: Option<ProgressCallback>,
+    pub(crate) write_parallelism: Option<usize>,
 }

 impl std::fmt::Debug for AddDataBuilder {
@@ -77,6 +82,8 @@ impl AddDataBuilder {
            write_options: WriteOptions::default(),
            on_nan_vectors: NaNVectorBehavior::default(),
            embedding_registry,
+            progress_callback: None,
+            write_parallelism: None,
        }
    }

@@ -101,7 +108,43 @@ impl AddDataBuilder {
        self
    }

+    /// Set a callback to receive progress updates during the add operation.
+    ///
+    /// The callback is invoked once per batch written, and once more with
+    /// [`WriteProgress::done`] set to `true` when the write completes.
+    ///
+    /// ```
+    /// # use lancedb::Table;
+    /// # async fn example(table: &Table) -> Result<(), Box<dyn std::error::Error>> {
+    /// let batch = arrow_array::record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+    /// table.add(batch)
+    ///     .progress(|p| println!("{}/{:?} rows", p.output_rows(), p.total_rows()))
+    ///     .execute()
+    ///     .await?;
+    /// # Ok(())
+    /// # }
+    /// ```
+    pub fn progress(mut self, callback: impl FnMut(&WriteProgress) + Send + 'static) -> Self {
+        self.progress_callback = Some(Arc::new(std::sync::Mutex::new(callback)));
+        self
+    }
+
+    /// Set the number of parallel write streams.
+    ///
+    /// By default, the number of streams is estimated from the data size.
+    /// Setting this to `1` disables parallel writes.
+    pub fn write_parallelism(mut self, parallelism: usize) -> Self {
+        self.write_parallelism = Some(parallelism);
+        self
+    }
+
    pub async fn execute(self) -> Result<AddResult> {
+        if self.write_parallelism.map(|p| p == 0).unwrap_or(false) {
+            return Err(Error::InvalidInput {
+                message: "write_parallelism must be greater than 0".to_string(),
+            });
+        }
+
        self.parent.clone().add(self).await
    }

@@ -130,8 +173,11 @@ impl AddDataBuilder {
            scannable_with_embeddings(self.data, table_def, self.embedding_registry.as_ref())?;

        let rescannable = self.data.rescannable();
+        let tracker = self
+            .progress_callback
+            .map(|cb| Arc::new(WriteProgressTracker::new(cb, self.data.num_rows())));
        let plan: Arc<dyn datafusion_physical_plan::ExecutionPlan> =
-            Arc::new(ScannableExec::new(self.data));
+            Arc::new(ScannableExec::new(self.data, tracker.clone()));
        // Skip casting when overwriting — the input schema replaces the table schema.
        let plan = if overwrite {
            plan
@@ -149,6 +195,7 @@ impl AddDataBuilder {
            rescannable,
            write_options: self.write_options,
            mode: self.mode,
+            tracker,
        })
    }
 }
@@ -161,6 +208,7 @@ pub struct PreprocessingOutput {
    pub rescannable: bool,
    pub write_options: WriteOptions,
    pub mode: AddDataMode,
+    pub tracker: Option<Arc<WriteProgressTracker>>,
 }

 /// Check that the input schema is valid for insert.
--- a/rust/lancedb/src/table/datafusion/insert.rs
+++ b/rust/lancedb/src/table/datafusion/insert.rs
@@ -12,13 +12,16 @@ use datafusion_common::{DataFusionError, Result as DataFusionResult};
 use datafusion_execution::{SendableRecordBatchStream, TaskContext};
 use datafusion_physical_expr::{EquivalenceProperties, Partitioning};
 use datafusion_physical_plan::execution_plan::{Boundedness, EmissionType};
+use datafusion_physical_plan::metrics::{ExecutionPlanMetricsSet, MetricBuilder, MetricsSet};
 use datafusion_physical_plan::stream::RecordBatchStreamAdapter;
 use datafusion_physical_plan::{
    DisplayAs, DisplayFormatType, ExecutionPlan, ExecutionPlanProperties, PlanProperties,
 };
+use futures::TryStreamExt;
 use lance::Dataset;
 use lance::dataset::transaction::{Operation, Transaction};
 use lance::dataset::{CommitBuilder, InsertBuilder, WriteParams};
+use lance::io::exec::utils::InstrumentedRecordBatchStreamAdapter;
 use lance_table::format::Fragment;

 use crate::table::dataset::DatasetConsistencyWrapper;
@@ -80,6 +83,7 @@ pub struct InsertExec {
    write_params: WriteParams,
    properties: PlanProperties,
    partial_transactions: Arc<Mutex<Vec<Transaction>>>,
+    metrics: ExecutionPlanMetricsSet,
 }

 impl InsertExec {
@@ -105,6 +109,7 @@ impl InsertExec {
            write_params,
            properties,
            partial_transactions: Arc::new(Mutex::new(Vec::with_capacity(num_partitions))),
+            metrics: ExecutionPlanMetricsSet::new(),
        }
    }
 }
@@ -176,6 +181,19 @@ impl ExecutionPlan for InsertExec {
        let total_partitions = self.input.output_partitioning().partition_count();
        let ds_wrapper = self.ds_wrapper.clone();

+        let output_bytes = MetricBuilder::new(&self.metrics).output_bytes(partition);
+        let input_schema = input_stream.schema();
+        let input_stream: SendableRecordBatchStream =
+            Box::pin(InstrumentedRecordBatchStreamAdapter::new(
+                input_schema,
+                input_stream.map_ok(move |batch| {
+                    output_bytes.add(batch.get_array_memory_size());
+                    batch
+                }),
+                partition,
+                &self.metrics,
+            ));
+
        let stream = futures::stream::once(async move {
            let transaction = InsertBuilder::new(dataset.clone())
                .with_params(&write_params)
@@ -215,6 +233,10 @@ impl ExecutionPlan for InsertExec {
            stream,
        )))
    }
+
+    fn metrics(&self) -> Option<MetricsSet> {
+        Some(self.metrics.clone_inner())
+    }
 }

 #[cfg(test)]
--- a/rust/lancedb/src/table/datafusion/scannable_exec.rs
+++ b/rust/lancedb/src/table/datafusion/scannable_exec.rs
@@ -7,17 +7,21 @@ use std::sync::{Arc, Mutex};
 use datafusion_common::{DataFusionError, Result as DFResult, Statistics, stats::Precision};
 use datafusion_execution::{SendableRecordBatchStream, TaskContext};
 use datafusion_physical_expr::{EquivalenceProperties, Partitioning};
+use datafusion_physical_plan::stream::RecordBatchStreamAdapter;
 use datafusion_physical_plan::{
    DisplayAs, DisplayFormatType, ExecutionPlan, PlanProperties, execution_plan::EmissionType,
 };
+use futures::TryStreamExt;

+use crate::table::write_progress::WriteProgressTracker;
 use crate::{arrow::SendableRecordBatchStreamExt, data::scannable::Scannable};

-pub struct ScannableExec {
-    // We don't require Scannable to by Sync, so we wrap it in a Mutex to allow safe concurrent access.
+pub(crate) struct ScannableExec {
+    // We don't require Scannable to be Sync, so we wrap it in a Mutex to allow safe concurrent access.
    source: Mutex<Box<dyn Scannable>>,
    num_rows: Option<usize>,
    properties: PlanProperties,
+    tracker: Option<Arc<WriteProgressTracker>>,
 }

 impl std::fmt::Debug for ScannableExec {
@@ -30,7 +34,7 @@ impl std::fmt::Debug for ScannableExec {
 }

 impl ScannableExec {
-    pub fn new(source: Box<dyn Scannable>) -> Self {
+    pub fn new(source: Box<dyn Scannable>, tracker: Option<Arc<WriteProgressTracker>>) -> Self {
        let schema = source.schema();
        let eq_properties = EquivalenceProperties::new(schema);
        let properties = PlanProperties::new(
@@ -46,6 +50,7 @@ impl ScannableExec {
            source,
            num_rows,
            properties,
+            tracker,
        }
    }
 }
@@ -102,7 +107,18 @@ impl ExecutionPlan for ScannableExec {
            Err(poison) => poison.into_inner().scan_as_stream(),
        };

-        Ok(stream.into_df_stream())
+        let tracker = self.tracker.clone();
+        let stream = stream.into_df_stream().map_ok(move |batch| {
+            if let Some(ref t) = tracker {
+                t.record_batch(batch.num_rows(), batch.get_array_memory_size());
+            }
+            batch
+        });
+
+        Ok(Box::pin(RecordBatchStreamAdapter::new(
+            self.schema(),
+            stream,
+        )))
    }

    fn partition_statistics(&self, _partition: Option<usize>) -> DFResult<Statistics> {
--- a/rust/lancedb/src/table/optimize.rs
+++ b/rust/lancedb/src/table/optimize.rs
@@ -64,6 +64,9 @@ pub enum OptimizeAction {
        older_than: Option<Duration>,
        /// Because they may be part of an in-progress transaction, files newer than 7 days old are not deleted by default.
        /// If you are sure that there are no in-progress transactions, then you can set this to True to delete all files older than `older_than`.
+        ///
+        /// **WARNING**: This should only be set to true if you can guarantee that no other process is
+        /// currently working on this dataset.  Otherwise the dataset could be put into a corrupted state.
        delete_unverified: Option<bool>,
        /// If true, an error will be returned if there are any old versions that are still tagged.
        error_if_tagged_old_versions: Option<bool>,
@@ -117,6 +120,10 @@ pub(crate) async fn optimize_indices(table: &NativeTable, options: &OptimizeOpti
 ///   If you are sure that there are no in-progress transactions, then you
 ///   can set this to True to delete all files older than `older_than`.
 ///
+///   **WARNING**: This should only be set to true if you can guarantee that
+///   no other process is currently working on this dataset.  Otherwise the
+///   dataset could be put into a corrupted state.
+///
 /// This calls into [lance::dataset::Dataset::cleanup_old_versions] and
 /// returns the result.
 pub(crate) async fn cleanup_old_versions(
--- a/rust/lancedb/src/table/query.rs
+++ b/rust/lancedb/src/table/query.rs
@@ -186,6 +186,13 @@ pub async fn create_plan(
        Select::Dynamic(ref select_with_transform) => {
            scanner.project_with_transform(select_with_transform.as_slice())?;
        }
+        Select::Expr(ref expr_pairs) => {
+            let sql_pairs: crate::Result<Vec<(String, String)>> = expr_pairs
+                .iter()
+                .map(|(name, expr)| expr_to_sql_string(expr).map(|sql| (name.clone(), sql)))
+                .collect();
+            scanner.project_with_transform(sql_pairs?.as_slice())?;
+        }
        Select::All => {}
    }

@@ -340,6 +347,17 @@ fn convert_to_namespace_query(query: &AnyQuery) -> Result<NsQueryTableRequest> {
                                .to_string(),
                    });
                }
+                Select::Expr(pairs) => {
+                    let sql_pairs: crate::Result<Vec<(String, String)>> = pairs
+                        .iter()
+                        .map(|(name, expr)| expr_to_sql_string(expr).map(|sql| (name.clone(), sql)))
+                        .collect();
+                    let sql_pairs = sql_pairs?;
+                    Some(Box::new(QueryTableRequestColumns {
+                        column_names: None,
+                        column_aliases: Some(sql_pairs.into_iter().collect()),
+                    }))
+                }
            };

            // Check for unsupported features
@@ -411,6 +429,17 @@ fn convert_to_namespace_query(query: &AnyQuery) -> Result<NsQueryTableRequest> {
                            .to_string(),
                    });
                }
+                Select::Expr(pairs) => {
+                    let sql_pairs: crate::Result<Vec<(String, String)>> = pairs
+                        .iter()
+                        .map(|(name, expr)| expr_to_sql_string(expr).map(|sql| (name.clone(), sql)))
+                        .collect();
+                    let sql_pairs = sql_pairs?;
+                    Some(Box::new(QueryTableRequestColumns {
+                        column_names: None,
+                        column_aliases: Some(sql_pairs.into_iter().collect()),
+                    }))
+                }
            };

            // Handle full text search if present
--- a/rust/lancedb/src/table/write_progress.rs
+++ b/rust/lancedb/src/table/write_progress.rs
@@ -0,0 +1,379 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+//! Progress monitoring for write operations.
+//!
+//! You can add a callback to process progress in [`crate::table::AddDataBuilder::progress`].
+//! [`WriteProgress`] is the struct passed to the callback.
+
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::{Arc, Mutex};
+use std::time::{Duration, Instant};
+
+/// Progress snapshot for a write operation.
+#[derive(Debug, Clone)]
+pub struct WriteProgress {
+    // These are private and only accessible via getters, to make it easy to add
+    // new fields without breaking existing callbacks.
+    elapsed: Duration,
+    output_rows: usize,
+    output_bytes: usize,
+    total_rows: Option<usize>,
+    active_tasks: usize,
+    total_tasks: usize,
+    done: bool,
+}
+
+impl WriteProgress {
+    /// Wall-clock time since monitoring started.
+    pub fn elapsed(&self) -> Duration {
+        self.elapsed
+    }
+
+    /// Number of rows written so far.
+    pub fn output_rows(&self) -> usize {
+        self.output_rows
+    }
+
+    /// Number of bytes written so far.
+    pub fn output_bytes(&self) -> usize {
+        self.output_bytes
+    }
+
+    /// Total rows expected.
+    ///
+    /// Populated when the input source reports a row count (e.g. a
+    /// [`arrow_array::RecordBatch`]).  Always `Some` when [`WriteProgress::done`]
+    /// is `true` — falling back to the actual number of rows written.
+    pub fn total_rows(&self) -> Option<usize> {
+        self.total_rows
+    }
+
+    /// Number of parallel write tasks currently in flight.
+    pub fn active_tasks(&self) -> usize {
+        self.active_tasks
+    }
+
+    /// Total number of parallel write tasks (i.e. the write parallelism).
+    pub fn total_tasks(&self) -> usize {
+        self.total_tasks
+    }
+
+    /// Whether the write operation has completed.
+    ///
+    /// The final callback always has `done = true`.  Callers can use this to
+    /// finalize progress bars or perform cleanup.
+    pub fn done(&self) -> bool {
+        self.done
+    }
+}
+
+/// Callback type for progress updates.
+///
+/// Callbacks are serialized by the tracker and are never invoked reentrantly,
+/// so `FnMut` is safe to use here.
+pub type ProgressCallback = Arc<Mutex<dyn FnMut(&WriteProgress) + Send>>;
+
+/// Tracks progress of a write operation and invokes a [`ProgressCallback`].
+///
+/// Call [`WriteProgressTracker::record_batch`] for each batch written.
+/// Call [`WriteProgressTracker::finish`] once after all data is written.
+///
+/// The callback is never invoked reentrantly: all state updates and callback
+/// invocations are serialized behind a single lock.
+impl std::fmt::Debug for WriteProgressTracker {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("WriteProgressTracker")
+            .field("total_rows", &self.total_rows)
+            .finish()
+    }
+}
+
+pub(crate) struct WriteProgressTracker {
+    rows_and_bytes: std::sync::Mutex<(usize, usize)>,
+    /// Wire bytes tracked separately by the insert layer. When set (> 0),
+    /// this takes precedence over the in-memory bytes from `rows_and_bytes`.
+    wire_bytes: AtomicUsize,
+    active_tasks: Arc<AtomicUsize>,
+    total_tasks: AtomicUsize,
+    start: Instant,
+    /// Known total rows from the input source, if available.
+    total_rows: Option<usize>,
+    callback: ProgressCallback,
+}
+
+impl WriteProgressTracker {
+    pub fn new(callback: ProgressCallback, total_rows: Option<usize>) -> Self {
+        Self {
+            rows_and_bytes: std::sync::Mutex::new((0, 0)),
+            wire_bytes: AtomicUsize::new(0),
+            active_tasks: Arc::new(AtomicUsize::new(0)),
+            total_tasks: AtomicUsize::new(1),
+            start: Instant::now(),
+            total_rows,
+            callback,
+        }
+    }
+
+    /// Set the total number of parallel write tasks (the write parallelism).
+    pub fn set_total_tasks(&self, n: usize) {
+        self.total_tasks.store(n, Ordering::Relaxed);
+    }
+
+    /// Increment the active task count. Returns a guard that decrements on drop.
+    pub fn track_task(&self) -> ActiveTaskGuard {
+        self.active_tasks.fetch_add(1, Ordering::Relaxed);
+        ActiveTaskGuard(self.active_tasks.clone())
+    }
+
+    /// Record a batch of rows passing through the scan node.
+    pub fn record_batch(&self, rows: usize, bytes: usize) {
+        // Lock order: callback first, then rows_and_bytes. This is the only
+        // order used anywhere, so deadlocks cannot occur.
+        let mut cb = self.callback.lock().unwrap();
+        let mut guard = self.rows_and_bytes.lock().unwrap();
+        guard.0 += rows;
+        guard.1 += bytes;
+        let progress = self.snapshot(guard.0, guard.1, false);
+        drop(guard);
+        cb(&progress);
+    }
+
+    /// Record wire bytes from the insert layer (e.g. IPC-encoded bytes for
+    /// remote writes). When wire bytes are recorded, they take precedence over
+    /// the in-memory Arrow bytes tracked by [`record_batch`].
+    pub fn record_bytes(&self, bytes: usize) {
+        self.wire_bytes.fetch_add(bytes, Ordering::Relaxed);
+    }
+
+    /// Emit the final progress callback indicating the write is complete.
+    ///
+    /// `total_rows` is always `Some` on the final callback: it uses the known
+    /// total if available, or falls back to the number of rows actually written.
+    pub fn finish(&self) {
+        let mut cb = self.callback.lock().unwrap();
+        let guard = self.rows_and_bytes.lock().unwrap();
+        let mut snap = self.snapshot(guard.0, guard.1, true);
+        snap.total_rows = Some(self.total_rows.unwrap_or(guard.0));
+        drop(guard);
+        cb(&snap);
+    }
+
+    fn snapshot(&self, rows: usize, in_memory_bytes: usize, done: bool) -> WriteProgress {
+        let wire = self.wire_bytes.load(Ordering::Relaxed);
+        // Prefer wire bytes (actual I/O size) when the insert layer is
+        // tracking them; fall back to in-memory Arrow size otherwise.
+        // TODO: for local writes, track actual bytes written by Lance
+        // instead of using in-memory Arrow size as a proxy.
+        let output_bytes = if wire > 0 { wire } else { in_memory_bytes };
+        WriteProgress {
+            elapsed: self.start.elapsed(),
+            output_rows: rows,
+            output_bytes,
+            total_rows: self.total_rows,
+            active_tasks: self.active_tasks.load(Ordering::Relaxed),
+            total_tasks: self.total_tasks.load(Ordering::Relaxed),
+            done,
+        }
+    }
+}
+
+/// RAII guard that decrements the active task count when dropped.
+pub(crate) struct ActiveTaskGuard(Arc<AtomicUsize>);
+
+impl Drop for ActiveTaskGuard {
+    fn drop(&mut self) {
+        self.0.fetch_sub(1, Ordering::Relaxed);
+    }
+}
+
+/// RAII guard that calls [`WriteProgressTracker::finish`] on drop.
+///
+/// This ensures the final `done=true` callback fires even if the write
+/// errors or the future is cancelled.
+pub(crate) struct FinishOnDrop(pub Option<Arc<WriteProgressTracker>>);
+
+impl Drop for FinishOnDrop {
+    fn drop(&mut self) {
+        if let Some(t) = self.0.take() {
+            t.finish();
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use std::sync::Arc;
+    use std::sync::atomic::{AtomicUsize, Ordering};
+
+    use arrow_array::record_batch;
+
+    use crate::connect;
+
+    #[tokio::test]
+    async fn test_progress_monitor_fires_callback() {
+        let db = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let table = db
+            .create_table("progress_test", batch)
+            .execute()
+            .await
+            .unwrap();
+
+        let callback_count = Arc::new(AtomicUsize::new(0));
+        let last_rows = Arc::new(AtomicUsize::new(0));
+        let max_active = Arc::new(AtomicUsize::new(0));
+        let last_total_tasks = Arc::new(AtomicUsize::new(0));
+        let cb_count = callback_count.clone();
+        let cb_rows = last_rows.clone();
+        let cb_active = max_active.clone();
+        let cb_total_tasks = last_total_tasks.clone();
+
+        let new_data = record_batch!(("id", Int32, [4, 5, 6])).unwrap();
+        table
+            .add(new_data)
+            .progress(move |p| {
+                cb_count.fetch_add(1, Ordering::SeqCst);
+                cb_rows.store(p.output_rows(), Ordering::SeqCst);
+                cb_active.fetch_max(p.active_tasks(), Ordering::SeqCst);
+                cb_total_tasks.store(p.total_tasks(), Ordering::SeqCst);
+            })
+            .execute()
+            .await
+            .unwrap();
+
+        assert_eq!(table.count_rows(None).await.unwrap(), 6);
+        assert!(callback_count.load(Ordering::SeqCst) >= 1);
+        // Progress tracks the newly inserted rows, not the total table size.
+        assert_eq!(last_rows.load(Ordering::SeqCst), 3);
+        // At least one callback should have seen an active task.
+        assert!(max_active.load(Ordering::SeqCst) >= 1);
+        // total_tasks should reflect the write parallelism.
+        assert!(last_total_tasks.load(Ordering::SeqCst) >= 1);
+    }
+
+    #[tokio::test]
+    async fn test_progress_done_fires_at_end() {
+        let db = connect("memory://").execute().await.unwrap();
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let table = db
+            .create_table("progress_done", batch)
+            .execute()
+            .await
+            .unwrap();
+
+        let seen_done = Arc::new(std::sync::Mutex::new(Vec::<bool>::new()));
+        let seen = seen_done.clone();
+
+        let new_data = record_batch!(("id", Int32, [4, 5, 6])).unwrap();
+        table
+            .add(new_data)
+            .progress(move |p| {
+                seen.lock().unwrap().push(p.done());
+            })
+            .execute()
+            .await
+            .unwrap();
+
+        let done_flags = seen_done.lock().unwrap();
+        assert!(!done_flags.is_empty(), "at least one callback must fire");
+        // Only the last callback should have done=true.
+        let last = *done_flags.last().unwrap();
+        assert!(last, "last callback must have done=true");
+        // All earlier callbacks should have done=false.
+        for &d in done_flags.iter().rev().skip(1) {
+            assert!(!d, "non-final callbacks must have done=false");
+        }
+    }
+
+    #[tokio::test]
+    async fn test_progress_total_rows_known() {
+        let db = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let table = db
+            .create_table("total_known", batch)
+            .execute()
+            .await
+            .unwrap();
+
+        let seen_total = Arc::new(std::sync::Mutex::new(Vec::new()));
+        let seen = seen_total.clone();
+
+        // RecordBatch implements Scannable with num_rows() -> Some(3)
+        let new_data = record_batch!(("id", Int32, [4, 5, 6])).unwrap();
+        table
+            .add(new_data)
+            .progress(move |p| {
+                seen.lock().unwrap().push(p.total_rows());
+            })
+            .execute()
+            .await
+            .unwrap();
+
+        let totals = seen_total.lock().unwrap();
+        // All callbacks (including done) should have total_rows = Some(3)
+        assert!(
+            totals.contains(&Some(3)),
+            "expected total_rows=Some(3) in at least one callback, got: {:?}",
+            *totals
+        );
+    }
+
+    #[tokio::test]
+    async fn test_progress_total_rows_unknown() {
+        use arrow_array::RecordBatchIterator;
+
+        let db = connect("memory://").execute().await.unwrap();
+
+        let batch = record_batch!(("id", Int32, [1, 2, 3])).unwrap();
+        let table = db
+            .create_table("total_unknown", batch)
+            .execute()
+            .await
+            .unwrap();
+
+        let seen_total = Arc::new(std::sync::Mutex::new(Vec::new()));
+        let seen = seen_total.clone();
+
+        // RecordBatchReader does not provide num_rows, so total_rows should be
+        // None in intermediate callbacks but always Some on the done callback.
+        let schema = arrow_schema::Schema::new(vec![arrow_schema::Field::new(
+            "id",
+            arrow_schema::DataType::Int32,
+            false,
+        )]);
+        let new_data: Box<dyn arrow_array::RecordBatchReader + Send> =
+            Box::new(RecordBatchIterator::new(
+                vec![Ok(record_batch!(("id", Int32, [4, 5, 6])).unwrap())],
+                Arc::new(schema),
+            ));
+        table
+            .add(new_data)
+            .progress(move |p| {
+                seen.lock().unwrap().push((p.total_rows(), p.done()));
+            })
+            .execute()
+            .await
+            .unwrap();
+
+        let entries = seen_total.lock().unwrap();
+        assert!(!entries.is_empty(), "at least one callback must fire");
+        for (total, done) in entries.iter() {
+            if *done {
+                assert!(
+                    total.is_some(),
+                    "done callback must have total_rows set, got: {:?}",
+                    total
+                );
+            } else {
+                assert_eq!(
+                    *total, None,
+                    "intermediate callback must have total_rows=None, got: {:?}",
+                    total
+                );
+            }
+        }
+    }
+}
Author	SHA1	Message	Date
Lance Release	f4d613565e	Bump version: 0.30.1 → 0.30.2-beta.0	2026-03-25 03:22:55 +00:00
Wyatt Alt	410ab9b6fe	Revert "feat: allow passing azure client/tenant ID through remote SDK" (#3185 ) Reverts lancedb/lancedb#3102	2026-03-24 20:17:40 -07:00
Will Jones	1d6e00b902	feat: progress bar for `add()` (#3067 ) ## Summary Adds progress reporting for `table.add()` so users can track large write operations. The progress callback is available in Rust, Python (sync and async), and through the PyO3 bindings. ### Usage Pass `progress=True` to get an automatic tqdm bar: ```python table.add(data, progress=True) # 100%\|██████████\| 1000000/1000000 [00:12<00:00, 82345 rows/s, 45.2 MB/s \| 4/4 workers] ``` Or pass a tqdm bar for more control: ```python from tqdm import tqdm with tqdm(unit=" rows") as pbar: table.add(data, progress=pbar) ``` Or use a callback for custom progress handling: ```python def on_progress(p): print(f"{p['output_rows']}/{p['total_rows']} rows, " f"{p['active_tasks']}/{p['total_tasks']} workers, " f"done={p['done']}") table.add(data, progress=on_progress) ``` In Rust: ```rust table.add(data) .progress(\|p\| println!("{}/{:?} rows", p.output_rows(), p.total_rows())) .execute() .await?; ``` ### Details - `WriteProgress` struct in Rust with getters for `elapsed`, `output_rows`, `output_bytes`, `total_rows`, `active_tasks`, `total_tasks`, and `done`. Fields are private behind getters so new fields can be added without breaking changes. - `WriteProgressTracker` tracks progress across parallel write tasks using a mutex for row/byte counts and atomics for active task counts. - Active task tracking uses an RAII guard pattern (`ActiveTaskGuard`) that increments on creation and decrements on drop. - For remote writes, `output_bytes` reflects IPC wire bytes rather than in-memory Arrow size. For local writes it uses in-memory Arrow size as a proxy (see TODO below). - tqdm postfix displays throughput (MB/s) and worker utilization (active/total). - The `done` callback always fires, even on error (via `FinishOnDrop`), so progress bars are always finalized. ### TODO - Track actual bytes written to disk for local tables. This requires Lance to expose a progress callback from its write path. See lance-format/lance#6247. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-23 16:14:13 -07:00
Esteban Gutierrez	a0228036ae	ci: fix unused PreprocessingOutput (#3180 ) Simple fix to for CI due unused import of PreprocessingOutput in table.rs Co-authored-by: Esteban Gutierrez <esteban@lancedb.com>	2026-03-23 13:45:44 -07:00
Esteban Gutierrez	d8fc071a7d	fix(ci): bump AWS SDK MSRV pins to March 2025 release (#3179 ) Lance v4.1.0-beta requires the default-https-client feature on aws-sdk-dynamodb and aws-sdk-s3, which was introduced in the March 2025 AWS SDK release. Update all AWS SDK pins to versions from the same AWS SDK release to maintain internal dependency compatibility. Co-authored-by: Esteban Gutierrez <esteban@lancedb.com>	2026-03-23 15:30:33 -05:00
Will Jones	e6fd8d071e	feat(rust): parallel inserts for remote tables via multipart write (#3071 ) Similar to https://github.com/lancedb/lancedb/pull/3062, we can write in parallel to remote tables if the input data source is large enough. We take advantage of new endpoints coming in server version 0.4.0, which allow writing data in multiple requests, and the committing at the end in a single request. To make testing easier, I also introduce a `write_parallelism` parameter. In the future, we can expose that in Python and NodeJS so users can manually specify the parallelism they get. Closes #2861 --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-20 13:19:07 -07:00
LanceDB Robot	670dcca551	feat: update lance dependency to v3.0.1 (#3168 ) ## Summary - Updated Lance Rust workspace dependencies to `3.0.1` using `ci/set_lance_version.py`. - Updated Java `lance-core` dependency property in `java/pom.xml` to `3.0.1`. - Refreshed `Cargo.lock` entries for Lance crates at `3.0.1`. ## Verification - `cargo clippy --workspace --tests --all-features -- -D warnings` - `cargo fmt --all` ## Trigger - Tag: [`refs/tags/v3.0.1`](https://github.com/lancedb/lance/tree/v3.0.1) Co-authored-by: Esteban Gutierrez <estebangtz@gmail.com>	2026-03-20 09:53:20 -07:00
Prashanth Rao	ed7e01a58b	docs: fix rendering issues with missing index types in API docs (#3143 ) ## Problem The generated Python API docs for `lancedb.table.IndexStatistics.index_type` were misleading because mkdocstrings renders that field’s type annotation directly, and the existing `Literal[...]` listed only a subset of the actual canonical SDK index type strings. Current (missing index types): <img width="823" height="83" alt="image" src="https://github.com/user-attachments/assets/f6f29fe3-4c16-4d00-a4e9-28a7cd6e19ec" /> ## Fix - Update the `IndexStatistics.index_type` annotation in `python/python/lancedb/table.py` to include the full supported set of canonical values, so the generated docs show all valid index_type strings inline. - Add a small regression test in `python/python/tests/test_index.py` to ensure the docs-facing annotation does not drift silently again in case we add a new index/quantization type in the future. - Bumps mkdocs and material theme versions to mkdocs 1.6 to allow access to more features like hooks After fix (all index types are included and tested for in the annotations): <img width="1017" height="93" alt="image" src="https://github.com/user-attachments/assets/66c74d5c-34b3-4b44-8173-3ee23e3648ac" />	2026-03-20 09:34:42 -07:00
Lance Release	3450ccaf7f	Bump version: 0.27.1-beta.0 → 0.27.1	2026-03-20 00:35:36 +00:00
Lance Release	9b229f1e7c	Bump version: 0.27.0 → 0.27.1-beta.0	2026-03-20 00:35:19 +00:00
Lance Release	f5b21c0aa4	Bump version: 0.30.1-beta.0 → 0.30.1	2026-03-20 00:35:03 +00:00
Lance Release	e927924d26	Bump version: 0.30.0 → 0.30.1-beta.0	2026-03-20 00:35:02 +00:00
Weston Pace	11a4966bfc	feat: upgrade lance dependency to v3.0.1 (#3157 ) ## Summary - Upgrade all lance-* dependencies from v3.0.0 to v3.0.1 (stable, from crates.io) ## Test plan - [x] `cargo check --features remote --tests --examples` passes - [x] `cargo clippy --features remote --tests --examples` passes - [x] `cargo fmt --all --check` passes - [ ] CI tests pass 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-19 17:30:46 -07:00
Weston Pace	dd5aaa72dc	ci: modify check_lance_release.py to prefer stable releases over betas (#3146 ) When Lance 3.0.0 released the check_lance_release.py script did not make a PR for it because it was a pre-release. This change may not be perfect but it always ranks stable releases above non-stable releases.	2026-03-17 09:21:30 -07:00
marca116	3a200d77ef	fix: pre-filtering on hybrid search (#3096 ) When using hybrid search with a where filter, the prefilter argument is silently inverted. Passing prefilter=True actually performs post-filtering, and prefilter=False actually performs pre-filtering.	2026-03-16 21:48:42 -07:00
Lance Release	bd09c53938	Bump version: 0.27.0-beta.6 → 0.27.0	2026-03-16 22:47:06 +00:00
Lance Release	0b18e33180	Bump version: 0.27.0-beta.5 → 0.27.0-beta.6	2026-03-16 22:46:48 +00:00
Lance Release	c89240b16c	Bump version: 0.30.0-beta.6 → 0.30.0	2026-03-16 22:46:19 +00:00
Lance Release	099ff355a4	Bump version: 0.30.0-beta.5 → 0.30.0-beta.6	2026-03-16 22:46:17 +00:00
Weston Pace	c5995fda67	feat: update lance dependency to 3.0.0 release (#3137 ) ## Summary - Update all 14 lance crates from `3.0.0-rc.3` (git source) to `3.0.0` (crates.io release) - Remove git/tag source references since 3.0.0 is published on crates.io ## Test plan - [x] `cargo check --features remote --tests --examples` passes - [x] `cargo clippy --features remote --tests --examples` passes - [ ] CI passes 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-16 15:29:18 -07:00
Weston Pace	25eb1fbfa4	fix: restore storage options on copy in localstack tests (#3148 )	2026-03-16 14:02:19 -07:00
Weston Pace	4ac41c5c3f	fix(ci): upgrade LocalStack to 4.0 for S3 integration tests (#3147 ) ## Summary - Upgrade LocalStack from 3.3 to 4.0 in `docker-compose.yml` to fix S3 integration test failures in CI - Version 3.3 has compatibility issues with newer Python 3.13 and updated boto3 dependencies - Matches the LocalStack version used successfully in the lance repository ## Test plan - [ ] Verify `docker compose up --detach --wait` completes successfully in CI - [ ] All tests in `test_s3.py` pass (5 tests) - [ ] All `@pytest.mark.s3_test` tests in `test_namespace_integration.py` pass (7 tests) - [ ] No regressions in non-integration test jobs (Mac, Windows) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-16 09:02:11 -07:00
Will Jones	9a5b0398ec	chore: fix ci (#3139 ) * Move away from buildjet, which is shutting down runners for GHA [^1] * Add `Cargo.lock` to build jobs, so when we upgrade locked dependencies we check the builds actually pass. CI started failing because dependencies were changed in #3116 without running all build jobs. * Add fixes for aws-lc-rs build in NodeJS. [^1]: https://buildjet.com/for-github-actions/blog/we-are-shutting-down --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-16 06:25:40 -07:00
Pratik Dey	d1d720d08a	feat(nodejs): support field/data type input in add_columns() method (#3114 ) Add support for passing field/data type information into add_columns() method, bringing parity with Python bindings. The method now accepts: - AddColumnsSql[] - SQL expressions (existing functionality) - Field - single Arrow field with explicit data type - Field[] - array of Arrow fields with explicit data types - Schema - Arrow schema with explicit data types New columns added via Field/Schema are initialized with null values. All field-based columns must be nullable due to null initialization. Resolves #3107 --------- Signed-off-by: Pratik <pratikrocks.dey11@gmail.com> Co-authored-by: Claude <noreply@anthropic.com>	2026-03-13 12:57:14 -07:00
Mesut-Doner	c2e543f1b7	feat(rust): support Expr in projection query (#3069 ) Referred and followed [`Select::Dynamic`] implementation. Closes #3039	2026-03-13 12:54:26 -07:00
Weston Pace	216c1b5f77	docs: remove experimental label from optimize and warn about delete_unverified (#3128 ) ## Summary - Removes the "Experimental API" section from `optimize` method documentation across Rust, Python, and TypeScript - Adds a warning to `delete_unverified` documentation in all bindings: this should only be set to true if you can guarantee no other process is working on the dataset, otherwise it could be corrupted - Fixes a typo ("shoudl" → "should") Closes #3125 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>	2026-03-12 14:37:42 +08:00