cshuo opened a new pull request, #20672: URL: https://github.com/apache/hudi/pull/20672
### Describe the issue this Pull Request addresses Closes #20668 MOR snapshot queries can miss records stored only in logs when metadata bloom data skipping is enabled. Base-file blooms do not cover those keys, but a negative result currently excludes the entire file slice. Log-only slices are also excluded because they have no candidate base filename. ### Summary and Changelog - Retain entire slices containing logs by including their log filenames in the candidates returned to `HoodieFileIndex`. - Skip base bloom lookups for slices with logs. Continue bloom pruning for base-only slices and read-optimized queries, whose input slices contain no logs. - Add a current-version MOR regression covering base+log and log-only records, snapshot equality/IN queries, missing keys, and read-optimized results. ### Impact Snapshot queries preserve matching records in logs. Slices containing logs are conservatively retained because their base blooms cannot rule out matches in the logs. ### Risk Level Low. The change is confined to Spark metadata bloom pruning and its regression test. The full `TestBloomFiltersIndexSupport` class passed on Spark 3.5.9 / Scala 2.12 / Java 11: 4 tests, no failures, errors, or skips. Checkstyle and Scalastyle passed. ### Documentation Update None. ### Contributor's checklist - [ ] 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]
