cshuo opened a new pull request, #20502: URL: https://github.com/apache/hudi/pull/20502
### Describe the issue this Pull Request addresses Closes #20501 Flink snapshot reads can still prune file slices with the record-level index when `read.data.skipping.enabled` is disabled, because the index receives the ungated filter probe. ### Summary and Changelog Use the probe gated by the data skipping and metadata settings when creating the record-level index. Add regression coverage with a real global record-level index for enabled and disabled data skipping, and verify that disabling metadata passes no evaluators to the index factory. ### Impact Queries with data skipping or metadata disabled no longer use record-level index pruning. These queries may scan more files; partition and bucket pruning remain available. ### Risk Level Low. The change reuses the existing data skipping gate and the record index factory's empty-evaluator fallback. The new regression and existing global and partitioned record index tests pass (35 tests): ```bash mvn -pl hudi-flink-datasource/hudi-flink -am -Pflink1.20 '-Dtest=TestFileIndex#testRecordLevelIndexRespectsDataSkipping+testFileListingWithRecordLevelIndex+testFileListingWithPartitionedRecordLevelIndex' -Dsurefire.failIfNoSpecifiedTests=false test ``` ### 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]
