zhangyue19921010 opened a new pull request, #11137: URL: https://github.com/apache/paimon/pull/11137
Data Evolution reads can fail with `StackOverflowError` when deletion vectors filter out many consecutive batches. `ForceSingleBatchReader` recursively calls `next()` to skip empty batches, causing stack usage to grow with their number. Replace the recursive call with a loop, preserving row order and batch release behavior. Tests cover consecutive empty batches before, between, and after records, fully empty input, and reader cleanup. A table-level regression deletes 19,999 of 20,000 rows and reads the remaining row across multiple column files with `read.batch-size=1`. Validation: the table regression fails on the unchanged baseline and passes with the fix. All 148 tests across seven related test classes pass with Maven `verify`. -- 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]
