JunRuiLee commented on PR #7832:
URL: https://github.com/apache/paimon/pull/7832#issuecomment-4544119616
> In my understanding, only when reading data should the sequenceNumbers in
the KeyValue be modified to the sequenceNumbers of the file, but the current
changes are much larger.
@JingsongLi Thanks for taking another look.
My understanding is that we may need to preserve the original
KeyValue.sequenceNumber as a secondary tiebreaker within the same snapshot. If
we replace it directly with the file-level sequence number on read, records
from the same snapshot would lose their original ordering.
Does it make sense to keep both values here: the file-level sequence
number as snapshotId, and the original KeyValue.sequenceNumber as the secondary
ordering key? Or do you think snapshot ordering can replace the original
sequence tiebreaker completely?
--
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]
