cshuo opened a new pull request, #20654: URL: https://github.com/apache/hudi/pull/20654
### Describe the issue this Pull Request addresses Closes https://github.com/apache/hudi/issues/20653 Flink primary-key predicates can discard native delete log records because the delete schema lacks the business-key columns. A snapshot query can then return an older base-file row that should have been deleted. ### Summary and Changelog - Skip query-predicate pushdown for native delete logs in `FlinkRowDataReaderContext`. - Preserve existing safe predicate pushdown for base files and data logs. - Add real Parquet regression tests for snapshot deletion, tombstone retention, and native data-log primary-key pruning. Validation: Flink 1.20.1 / Java 11; all 67 tests passed across `TestFlinkNativeDeleteLogFilters`, `TestFlinkRowDataReaderContext`, `TestRecordIterators`, `TestExpressionPredicates`, and `TestHoodieFileGroupReaderNativeLogs`. Checkstyle passed. ### Impact Corrects Flink snapshot results when primary-key filters are used with native delete logs. Native delete logs are read without query predicates; base/data-file filtering is unchanged. No storage-format or public API changes. ### Risk Level Low. The change is limited to native delete log predicate selection and is covered by file-group reader integration tests using actual Parquet files. ### 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]
