XiaoHongbo-Hope opened a new pull request, #8981: URL: https://github.com/apache/paimon/pull/8981
Stacked on #8980, #8979, and #8826. The effective diff after those PRs is seven files. This adds `PaimonCoBucketedJoinOffsetSource` for an end-to-end resumable pipeline: - pin and co-bucket join two Paimon source snapshots; - route matched row ids once into an internal table bucketed by target file group; - checkpoint route construction and target-group update windows with bounded offsets; - resume without recomputing completed transform/update windows; - reject external target rewrites and transforms that move row ids across windows; - drop routing state and retained tags through `delete_update_by_row_id_checkpoint`. Tests: `ray_update_by_row_id_test.py` (36 passed); read/merge suites (82 passed, 26 skipped) with Ray 2.56.1; project flake8 clean. -- 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]
