SEZ9 commented on issue #13244:
URL: https://github.com/apache/seatunnel/issues/13244#issuecomment-6845165900

   Thanks for the detailed report. Both points look like real correctness 
issues: (1) `write()` turns every row into a `Put` with no `RowKind` check, so 
`DELETE` and `UPDATE_BEFORE` rows should be handled differently instead of 
being written back as normal rows, and (2) `prepareCommit()` should flush the 
`BufferedMutator` so that rows acknowledged by a checkpoint are actually 
durable in HBase rather than only being flushed in `close()`.
   
   Since you are willing to submit a PR, that would be great. A 
FakeSource-based e2e test covering `INSERT`, `UPDATE_BEFORE`, `UPDATE_AFTER` 
and `DELETE` rows would be a good regression test for the first defect. Could 
you share the exact FakeSource config you used for the batch reproduction so it 
can be turned into that test case?
   
   <!-- streview-comment:2323 -->


-- 
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]

Reply via email to