Xuanwo
725eee8a47
Ambiguous gateway failures can occur after a write commits, so replaying the operation can append duplicate rows. Non-idempotent writes now retry only known precommit failures. Multipart retries reuse the same upload ID, part ID, and encoded bytes; ambiguous completion returns an unknown-outcome error without abort or replay. A definite 409/429 completion rejection triggers best-effort cleanup using the call's retry budget. Replayable parts and the shared byte-based partition target default to 64 MiB. Dictionary-bearing batches are compacted before size estimation and again when actual compressed IPC framing requires further splitting: keys are remapped without changing their types or schema, view backing buffers are collected, and ListView/LargeListView child ranges are compacted as an interval union so overlapping views retain sharing. Parts are prepared on demand; a single row and its schema can still exceed the target. Multipart progress starts at zero and counts each acknowledged IPC body once. Rust write builders and Python/TypeScript `add` expose per-call retry overrides. Atomic `create_table` pre-encodes to a temporary file and rejects compressed IPC over 256 MiB before sending any request; the schema-first import guidance now includes Python, TypeScript, and Rust examples. The shared partition target also affects local automatic partitioning. At `cb2a79473bd1b51b950c9fa3e9f247f929c8ac3b`, 70 targeted Rust cases pass on source verified against the committed four-file patch. With a 16 KiB target and 128 unique 1 KiB dictionary values, the original ordinary dictionary reproducer drops from 128 requests / 16,942,080 bytes to 16 / 143,488; ListView and LargeListView drop from 128 / 17,056,768 to 16 / 149,632. Encoded HTTP bodies are decoded to check values, nulls, slices, metadata, changing dictionaries, and overlapping views; additional cases cover framing-driven splitting, oversized rows, stable retry bytes/IDs, acknowledged progress, and atomic create. Real HTTP fault tests commit Blob payloads to a local Lance table and verify byte equality, no duplicate append, and zero abort/replay after ambiguous completion. The earlier 811 remote and 7 progress cases at `f36d217110a25f6b501c0255df9aaa60e3c8000a`, Linux Python remote database/options results (437 cases plus 8 fixture consumers), Node retry-binding evidence, and Rust/TypeScript documentation example checks remain historical validation. Python/Node extensions were not rebuilt for the dictionary follow-up; these results do not establish current binding integration or fresh cross-platform/minimum-dependency CI acceptance. A separate real Query Node lifecycle experiment used SDK `b8efe7d2d63647d4962789ca5fc042b59efe073d` and Sophon `58b7895c7c287810fff4e89f13226fc5fa60d9ef`. After the actual listener forced a connection closed, a surviving QN accepted a retry with identical upload ID, part ID, and IPC bytes. Ordinary insert and complete response-loss scenarios sent no replay or abort; fresh reads observed exactly one commit and exact Blob bytes in all three scenarios. This local experiment used a 1-second drain grace, two 32 KiB payloads, and one distinct staged part. It validates acknowledgement-loss recovery at those revisions, not concurrent overlapping attempts, large uploads, physical orphan cleanup, the later dictionary encoding change, or the original sixth-insert EOF. The part target is not a process memory limit: source/ancestor buffers and temporary compaction allocations may coexist. Struct/FixedSizeList parent nulls can retain hidden child keys; this is not complete logical-reachability GC for every Arrow combination, and retained hidden values may affect wire size. The current local Lance 15.0.0-beta.17 rejects ListView/LargeListView persistence; the tests establish remote client IPC encoding behavior, not native or Cloud persistence support. Live Cloud, full workspace/Clippy, the original ordinary-insert disconnect root cause, and Sophon's asynchronous transaction cleanup behavior are not established. Cancelling during staging retains the existing server-session cleanup limitation. Function reader delivery is separate.
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"])