SEZ9 commented on PR #12608: URL: https://github.com/apache/seatunnel/pull/12608#issuecomment-6050856818
Thanks @CodeWithPravinMaske for the follow-up and the links. **Docs / incompatible-changes.** Thanks for pointing me to the MySQL-CDC "Read tables without a primary key" section and `incompatible-changes.md` (en/zh) at `fa70f3eb8`. I'll re-check the diff on my side; if the docs now describe NULL being emitted as NULL for a nullable column configured in `table-names-config.primaryKeys`, could you also point me to where that path is handled in `MySqlSchema`, so the doc statement and the code line up? **Binlog / checkpointed DDL path.** Good to see `testMysqlCdcNullInNullableUniqueKeyAfterRestore` passing in run `37420097470` (jobs `112129448460` and `112129448548`). A restore test alone doesn't tell me whether the nullability restore is also applied to DDL parsed from the binlog (`parseStreamingDdl`) and from the checkpointed table history. Could you point me to the code change covering those paths, and, if there is one, a test that applies a unique-key DDL change during the streaming phase? **Remaining smaller items.** Could you confirm (ideally with the relevant diff) the status of: - the `catalogTable == null` guard / "may be null" Javadoc in `restoreNullableColumns`; - the per-column `log.info` in the restore (moved to `debug` or aggregated per table?). I'll do another pass once those are in. <!-- streview-comment:1585 --> -- 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]
