SteNicholas opened a new pull request, #11118:
URL: https://github.com/apache/paimon/pull/11118

   ### Purpose
   
   Fixes #11117.
   
   On data evolution tables, commit replaces a file's sequence range with the 
snapshot id only when its min sequence number is 0. PyPaimon's `BlobWriter` 
numbered its rolled files by row, so every blob file after the first kept 
row-based sequence numbers. These can equal or exceed the snapshot id of a 
later update-by-row-id, so reads failed with `Blob files within the same max 
sequence should not overlap` or returned the old blob.
   
   Like Java (`AppendOnlyWriter` provides a new `LongCounter(0)` per rolled 
file on data evolution tables), `BlobWriter` now gives every blob file the 
sequence range `(0, row_count - 1)` on data evolution tables, so commit assigns 
the snapshot id to all of them.
   
   ### Tests
   
   - Added 
`DedicatedFormatWriterTest#test_update_blob_column_after_rolling_write`: writes 
10 rows with `blob.target-file-size = 1 B` (one blob file per row), checks that 
every blob file carries the commit's snapshot id, then updates rows 2 and 7 by 
row id and checks the read result. It fails without the fix.
   
   🤖 Generated with [Claude Code](https://claude.com/claude-code)
   


-- 
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