danny0405 opened a new pull request, #20340:
URL: https://github.com/apache/hudi/pull/20340
### Describe the issue this Pull Request addresses
Following #20273, `CompletionTimeQueryViewV2.isArchived()` excludes
savepointed instants but still classifies active clean/rollback instants before
the first non-savepointed write as archived. If their completion times fall in
an incremental query's range, `IncrementalQueryAnalyzer` unnecessarily loads
the archived timeline. This can occur even when archival beyond savepoints is
disabled.
### Summary and Changelog
Use membership in the meta client's cached active timeline to exclude all
active instants from archive classification.
- Replace the separate savepoint timestamp set with
`getActiveTimeline().containsInstant()` while preserving the existing write
boundary and lazy-loading cursor.
- Clarify that tables upgraded from pre-v8 can have interleaved archived and
active instants.
- Extend existing tests with clean/rollback cases, with and without
savepoints, verifying that incremental queries avoid archive lookups and that
archived writes remain readable around older active instants.
No code was copied from external projects.
### Impact
Avoids unnecessary archived-timeline initialization and manifest lookup for
incremental queries containing active clean/rollback completion times. When
those instants are newer than all archived files, existing filename-range
filtering already skips the Parquet files, so the benefit is limited to setup
and metadata overhead. Older interleaved active instants can overlap archive
file ranges. The membership check reuses the cached active timeline and adds no
storage I/O. No public API, configuration, or storage-format changes.
### Risk Level
low
The change is limited to archive classification in the V2 completion-time
query view. The existing cursor and archive-loading behavior are preserved. All
21 targeted tests passed across `TestIncrementalQueryAnalyzer`,
`TestCompletionTimeQueryView`, and `TestCompletionTimeQueryViewV1`, including
savepoint and interleaving coverage.
Validation used Java 17 with the Byte Buddy agent preloaded:
```sh
mvn -o -pl hudi-client/hudi-client-common -am test \
-Dtest=TestCompletionTimeQueryView,TestCompletionTimeQueryViewV1,TestIncrementalQueryAnalyzer
\
-Dsurefire.failIfNoSpecifiedTests=false -DfailIfNoTests=false \
-DskipITs -DskipFTs -Drat.skip=true
```
Integration/functional test suites and RAT were not run. `git diff --check`
passed.
### Documentation Update
none — internal query optimization with no new feature or configuration. The
constructor comment explains the pre-v8 upgrade case.
### Contributor's checklist
- [x] Read through [contributor's
guide](https://hudi.apache.org/contribute/how-to-contribute)
- [x] Enough context is provided in the sections above
- [x] Adequate tests were added if applicable
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
