SEZ9 commented on issue #10451: URL: https://github.com/apache/seatunnel/issues/10451#issuecomment-6159392322
@wgzhao Thanks for tracing this down to the derived `unsigned` flag — that corrects my earlier description. The `sql_mode` columns are plain `SET` columns; the `UNSIGNED` came from the old `COLUMN_TYPE` substring match hitting the value name `NO_UNSIGNED_SUBTRACTION`, not from a real `SET UNSIGNED` type. To answer your question directly: no, please do not add the `MYSQL_SET_UNSIGNED` / `ENUM UNSIGNED` cases, neither as the fix nor as a separate belt-and-braces change. It would map an invented type to `STRING` while leaving the column flagged `unsigned`, and it is unreachable once the derivation guard stops setting the flag for types that cannot carry it. My earlier "approach 1" was based on the wrong premise, so I withdraw it. Where this leaves us: - Your derivation-fix PR (the one marked as fixing this issue) is the resolution for this report. This issue stays open until it is merged and its validation has completed. - The narrowed `database-pattern` remains the workaround for released versions; it only avoids scanning `mysql.*`, it does not fix the derivation. - The follow-up cleanup you opened (guard derived from the converter constants, `DATA_TYPE`-based check, OceanBase E2E case) is fine as a separate stacked change. Nothing else needed from you on this thread right now — I will follow up on the PR itself. <!-- streview-comment:1731 --> -- 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]
