Yang Cen
69ac2158a1
`take_row_ids` used an `IN` filter, so the backend could reorder rows and collapse duplicate IDs. That can silently misalign training rows with labels or sample weights. Restore the requested order and duplicate occurrences in the shared Rust take path for native and remote tables, and retain that information through Python query serialization. Missing or deleted IDs are omitted. Explicit Rust `order_by` overrides request order, and pagination applies after restoring occurrences. Empty projections and `_rowid` projection aliases retain their output schema. RTE tables reject unsupported takes with the existing base-table guidance before checking the internal row-ID projection used for ordering. An explicit `use_rte(False)` reads the base table with the same requested order and occurrences. A pre-existing fixture defect also blocked CI: `test_table_add_in_threadpool` responded without reading uploads, allowing a connection reset during response decoding. This PR includes a fix confined to that test's handler: consume each request body before responding and verify one create row plus ten inserted rows. This was not a regression in the take implementation. Production write retries, concurrency and HTTP client code are unchanged. The selective sampling guide scopes the row-ID guarantees to SDK releases containing this fix, preserves the distinct offset-take semantics, and explains that ordered takes collect and combine matching rows before output batching. It recommends smaller ID requests and narrow projections because `max_batch_length` does not cap working memory. This uses the existing REST query protocol and requires a new SDK release; no deployed service was changed. Validation: - Docs follow-up: the changed page compiles with the Mint 4.2.888 MDX compiler; Typos 1.26.0 passes and Vale 3.24.0 has no findings on changed lines. Full-site Mint validation reports four existing snippet-import warnings on the reranking/vector-search pages, identical with the original page restored in a baseline comparison. - The applied threadpool test passed locally after the fixture change. The identical candidate previously passed 10 trials with the ordinary reader and 10 with an unbuffered reader; each trial verifies all uploaded rows. Those checks used macOS arm64/Python 3.14/default features, not the full failing CI environment. The normal new-head macOS Python 3.13/experimental-functions job remains the required platform check. - Python package Ruff format/check passed. Repository-root Ruff still reports 59 existing errors and six formatting differences in unchanged `docs/web-tests` files; none are in changed files. - At `a569d3f0`, the RTE diagnostic regression was reproduced and fixed; 5 targeted cases and strict workspace Clippy passed. The fixture-only follow-up changes no production Rust/Python code, so that evidence is reused. - Earlier validation at `6bcc0878`: 439 Rust query/remote-table tests and 24 Python take/getitems tests passed; the focused 49-test suite also covered offset dataloaders. A local 9,216-row, four-fragment reproduction returned 200 seeded random batches of 64 rows with zero ordering or label-alignment mismatches. The ordering and projection implementation is unchanged by the follow-ups. - No live Enterprise run was performed. Fixes [ENT-3469](https://linear.app/lancedb/issue/ENT-3469).
LanceDB Python SDK
A Python library for LanceDB.
Installation
pip install lancedb
If no wheel is published for your platform, pip builds lancedb from the source distribution. That build needs a Rust toolchain and protoc (the Protocol Buffers compiler) on your PATH, and can take a while.
Pre-Haswell x86_64 hosts: lancedb-compat
The default lancedb wheel targets x86-64-haswell (AVX2 + FMA + F16C) for full performance on modern hardware. Pre-Haswell hosts — Intel Sandy Bridge / Ivy Bridge / Westmere; AMD Bulldozer / Piledriver / Steamroller — don't have AVX2 and crash with Illegal instruction at import lancedb.
For those hosts, install the lancedb-compat package instead:
pip install lancedb-compat
Same Python API (import lancedb works as usual). The compat wheel is compiled at the x86-64-v2 baseline (Nehalem-class) and uses runtime SIMD dispatch in the embedded lance crate to pick the right kernel tier (scalar / AVX / AVX+FMA / AVX2+FMA / AVX-512) at load time, so it still goes fast on modern hardware while running cleanly on the pre-Haswell silicon (if pylance is installed, lance.simd_info() reports which tier was selected).
lancedb and lancedb-compat install to the same lancedb/ namespace and conflict at install time. Pick one. To switch, pip uninstall lancedb first, then pip install lancedb-compat (or vice-versa).
If you need a custom baseline (or lancedb-compat isn't yet published for your platform), build from source with the override:
RUSTFLAGS="-C target-cpu=x86-64-v2" maturin build --release
pip install ./target/wheels/lancedb-*.whl
Preview Releases
Stable releases are created about every 2 weeks. For the latest features and bug fixes, you can install the preview release. These releases receive the same level of testing as stable releases, but are not guaranteed to be available for more than 6 months after they are released. Once your application is stable, we recommend switching to stable releases.
pip install --pre --extra-index-url https://pypi.fury.io/lancedb/ lancedb
Most applications only need lancedb. Install the optional pylance extra and
include the Lance preview index only when building query engines or connectors,
calling format-level lance APIs directly, or converting plain-scan lazy or
legacy blob columns with to_pandas():
pip install --pre \
--extra-index-url https://pypi.fury.io/lancedb/ \
--extra-index-url https://pypi.fury.io/lance-format/ \
"lancedb[pylance]"
With uv, also select the best matching version across the indexes:
uv pip install --prerelease allow \
--index https://pypi.fury.io/lancedb/ \
--index https://pypi.fury.io/lance-format/ \
--index-strategy unsafe-best-match \
"lancedb[pylance]"
The LanceDB preview index contains old pylance releases; current Lance previews
are published to the lance-format index. By default, uv only considers versions
from the first index containing a package, which can select an old pylance or
backtrack to an older LanceDB release. unsafe-best-match searches all configured
indexes, like pip; use it only with indexes you trust. Install the extra rather
than an unconstrained pylance so LanceDB's minimum supported version is enforced.
Threading in CPU-limited containers
LanceDB uses separate pools for compute work and storage I/O. On a container with two visible CPUs, current releases intentionally use one compute worker by default; no manual configuration is needed. If every query logs an I/O core reservation warning on a two-CPU container, upgrade from LanceDB 0.21.1 or earlier.
The two commonly tuned environment variables control different resources:
LANCE_CPU_THREADSoverrides the number of compute workers. One worker is the appropriate setting for a two-CPU container when an explicit override is needed.LANCE_IO_THREADScontrols concurrent storage operations, not reserved CPU cores. Its default can be greater than the number of CPUs because I/O workers spend much of their time waiting for storage.
Keep the defaults unless measurements show that the workload benefits from an override. See the Lance threading model for the current defaults and tuning guidance.
Usage
Basic Example
import lancedb
db = lancedb.connect('<PATH_TO_LANCEDB_DATASET>')
table = db.open_table('my_table')
results = table.search([0.1, 0.3]).limit(20).to_list()
print(results)
Development
See CONTRIBUTING.md for information on how to contribute to LanceDB.
Opening SQL blob descriptors
A SQL blob_descriptor(video) result keeps the Blob V2 storage metadata and a
source struct containing its table URI, version, commit timestamp, branch,
column and row identity. This example uses a remote SQL connection whose SQL
service produces those source-bearing descriptors. Open the reference on the
source table to read incrementally:
rows = db.execute_query("""
SELECT f.timestamp, blob_descriptor(v.video) AS video
FROM videos v JOIN frames f ON v.id = f.video_id
""").read_all()
videos = db.open_table("videos")
for row in rows.to_pylist():
if row["video"] is not None:
with videos.open_blob_file(row["video"]) as video:
frame = extract_frame(video, row["timestamp"])
open_blob_file accepts the dictionary or an Arrow struct scalar directly. It
keeps the query's version and branch without changing the table handle's checkout.
Payload bytes load on reads. An external blob with an unspecified size may need a
metadata-only HEAD request to open.
videos.open_blob_files(rows["video"]) returns an iterator over the same handles,
preserving order, duplicates and nulls (None). It accepts Arrow arrays/chunked
arrays or another iterable and opens each handle as iteration advances. Close each
non-null handle after use. Async tables offer await videos.open_blob_file(...)
and async for video in videos.open_blob_files(...).
for video in videos.open_blob_files(rows["video"]):
if video is not None:
with video:
header = video.read(4096)
To extract a frame while retaining each row's timestamp, iterate over the handles and rows together:
for row, video in zip(rows.to_pylist(), videos.open_blob_files(rows["video"])):
if video is not None:
with video:
frame = extract_frame(video, row["timestamp"])