Skip to content

[flink] Add local-sort mode for incremental clustering compaction#7466

Open
lilei1128 wants to merge 2 commits intoapache:masterfrom
lilei1128:feature
Open

[flink] Add local-sort mode for incremental clustering compaction#7466
lilei1128 wants to merge 2 commits intoapache:masterfrom
lilei1128:feature

Conversation

@lilei1128
Copy link
Contributor

Purpose

Add clustering.incremental.mode option with two values:

  • global-sort (default): existing behavior, global range shuffle + local sort
  • local-sort: sort rows only within each compaction task without global shuffle,
    producing files that are internally ordered by clustering columns

The local-sort mode is cheaper (no network shuffle) and sufficient for
per-file Parquet lookup optimizations (PIP-42), where only intra-file
ordering is required.

Linked issue: close #7253

Tests

CI

Documentation

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Feature] Support order data in single file for append table

1 participant