Repository navigation
[core] Support row ID reassignment across concurrent compaction - #10780
Merged
JingsongLi merged 9 commits intoOct 26, 2026
Merged
JingsongLi merged 9 commits into
JingsongLi merged 9 commits into
Conversation
leaves12138
marked this pull request as ready for review
October 26, 2026 06:21
Contributor
|
+1 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Purpose
Row ID reassignment currently aborts on every concurrent COMPACT snapshot. Allow it to reuse existing mappings and bind rewrites to the current manifests: rewrite mapped ADD/DELETE ranges, leave disjoint ranges unchanged, and reject partially mapped ranges. Initial planning and commit retries keep their existing behavior.
Concurrent deletion-vector materialization can retain new-file row IDs while historical ADD/DELETE pairs are remapped consistently. Removed ranges may leave unused row IDs in the original assignment.
Validation
Passed 110 checked-in tests on Linux / OpenJDK 11 with normal Maven checks:
mvn -pl paimon-core -DwildcardSuites=none \ -Dtest=DataEvolutionRowIdReassignerTest,CompactReassignReuseTest,DataEvolutionDeletionVectorTest testReuse existing retry tests for data/manifest compaction and partial-overlap rejection. Add one focused DV materialization test covering BLOB values, partial-column updates, unchanged new-file IDs and surviving old-file mappings. Related tests contain no reflection.
Additional isolated validation passed 20 temporary compaction/DV cases and 26 metadata replays (24 partition slices and two full snapshots), including independent range audits and 148,616,799 full-snapshot entry comparisons. Eight negative cases were rejected as expected, including missing/expired metadata.