mirror of
https://github.com/lancedb/lancedb.git
synced 2026-11-13 22:26:53 +00:00
Jack YeandClaude Sonnet 5
de95ed3f6d
A broad documentation pass across the areas LanceDB Enterprise and Embedded diverge on most: **Data Curation** -- GPU Function execution, Slurm scheduling (and what Slurm is, for readers new to it), blob late materialization as its own page, joins, aggregation/profiling, sampling, near-duplicate detection, chunking, distributed shuffle, and IVF-partition clustering. **Feature Engineering** -- computed columns and materialized views, restructured to lead with the Function-based path and show the SQL form side by side with Python throughout; a single Auto Refresh page instead of duplicating scheduling details on every refresh page; adaptive execution; checkpointing; GPU and Slurm placement for refresh workers; a Training Cache page for the PyTorch data loader's automatic cache benefit. **Security & Governance** -- an Audit Log reference (schema, access, grants), Secrets promoted to an ordinary catalog object with their own SQL DDL and a vending/rotation guide, and privilege-table fixes verified against the authorization source. **Deployment & Maintenance** -- Horizontal Scaling and OpenTelemetry moved into a real Deployment area; Automated Cleanup folded into Intelligent Auto-Maintenance (every other maintenance decision already lived there); RTE Compaction and Advanced Conflict Resolution documented in depth. **Landing page** -- the Enterprise feature comparison table reorganized (several redundant or stale rows removed, several new areas added), and the "why Enterprise" bullets replaced with ones tied to a concrete page (Performance, Scale, Fault tolerance, Multicloud deployment, Simplify operations) instead of vague promises. Plus a full pass addressing lancedb-gatekeeper's review: a real test-safety bug (a quickstart test could silently overwrite and delete a deployment's own `frames` table), a fabricated SQL statement with no backing implementation, an actual dedup-guarantee bug in `take_row_ids`, and dozens of documentation-accuracy fixes verified against source rather than taken on faith -- each either fixed, or left with an explicit note on why it needs a feature or deployment owner's sign-off instead of being guessed at. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_018PXMCbRH6i5w47TsjeEUMT --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
52 lines
2.3 KiB
Plaintext
52 lines
2.3 KiB
Plaintext
---
|
|
title: "Incremental Refresh"
|
|
sidebarTitle: "Incremental Refresh"
|
|
description: "Why refreshing a computed column only recomputes fragments that hold a row whose inputs actually changed."
|
|
icon: "arrow-rotate-right"
|
|
availability:
|
|
embedded: unavailable
|
|
enterprise: available
|
|
---
|
|
|
|
Refreshing a [computed column](/feature-engineering/add-computed-column)
|
|
is incremental: it fills rows that have no value yet and recomputes the
|
|
fragments that hold a row whose input changed since it was last computed,
|
|
leaving every other fragment alone. A second refresh with nothing changed
|
|
does nothing. The
|
|
planner finds the rows that actually need work by checking which
|
|
fragments still carry stale or empty values for the column, then groups
|
|
them into claimable units a fleet of job workers can pick up in parallel
|
|
-- see [Architecture](/index#architecture).
|
|
|
|
<Note>
|
|
A **materialized view** doesn't track stale rows the way a computed
|
|
column does -- there's no smaller unit than the whole view to recompute
|
|
from a saved query. In LanceDB Embedded, every refresh rebuilds the view
|
|
completely with no incremental step; LanceDB Enterprise's server chooses
|
|
the refresh mode instead, unless the caller forces a full rebuild. See
|
|
[Refresh a materialized view](/feature-engineering/refresh-materialized-view).
|
|
</Note>
|
|
|
|
## Why this matters {#why}
|
|
|
|
Appending rows is cheap: the next refresh only fills what's new, not what
|
|
it already computed. Updating an input row is also cheap in the same
|
|
way -- the derived column keeps its old value until the next refresh
|
|
notices the input changed. Staleness is tracked per fragment, not per
|
|
row: the refresh recomputes every row in a fragment where at least one
|
|
input moved, which can re-evaluate rows whose own value didn't actually
|
|
need it, not strictly only the row that changed. Nothing is recomputed
|
|
between refreshes, so a derived column can be stale right up until you
|
|
call for one.
|
|
|
|
## What to read next {#next}
|
|
|
|
<CardGroup cols={2}>
|
|
<Card title="Refresh a computed column" icon="arrows-rotate" href="/feature-engineering/refresh-column#idempotent">
|
|
The exact fill/recompute/leave-alone rules, with examples.
|
|
</Card>
|
|
<Card title="Adaptive Execution" icon="gauge-high" href="/enterprise/adaptive-execution">
|
|
How the units this planner produces get sized and run.
|
|
</Card>
|
|
</CardGroup>
|