SEZ9 commented on PR #11746: URL: https://github.com/apache/seatunnel/pull/11746#issuecomment-6228596481
@li3zhi4 thanks for part 2/2 on head `e7586c06a83ae174775bd3fdd8aa46a05039672f`. **F5 — comma separator:** accepted. Since the split carries key names/types and the boundary arrays structurally, values are never encoded into the joined name, and nothing decodes on the separator, the comma issue reduces to an ambiguous display/metadata name, which the Javadoc you linked calls out explicitly. I'm fine treating F5 as a documented limitation rather than a bug; no escaping codec or dedicated regression needed. **F6–F8:** your comment appears cut off right at the start of the F6 link, so I only have the F5 portion. Could you re-post the rest? Specifically: - **F6:** where the element/numeric comparison normalizes mixed JDBC numeric types (e.g. SQLite returning Integer vs Long for the same column) and how BigDecimal is compared so scale differences don't defeat the dedup/equality guard. - **F7:** where the upgrade note lives in the JDBC source connector docs and whether there is a documented way to opt back into the old single-column split for existing composite-PK jobs. - **F8:** what now protects `FixedChunkSplitter` from a multi-column split key after the shared guard was removed from `ChunkSplitter.generateSplits`. Once I have those pointers I'll confirm F6–F8 the same way as F5. <!-- streview-comment:1810 --> -- 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]
