XiaoHongbo-Hope opened a new pull request, #8985:
URL: https://github.com/apache/paimon/pull/8985

   ### Purpose
   
   PyPaimon batch writes to postpone-bucket tables currently always create 
`bucket-postpone` files. Unlike the Java batch integrations, those files are 
not visible to normal readers until compaction finishes.
   
   This aligns PyPaimon batch writes with the default 
`postpone.batch-write-fixed-bucket=true` behavior so committed rows are 
immediately visible.
   
   ### Changes
   
   - Route PyPaimon batch rows directly to real buckets for `bucket=-2` tables.
   - Reuse the current bucket count for existing partitions and use 
`postpone.default-bucket-num` (default `1`) for new partitions.
   - Carry the per-partition total bucket count through commit messages and 
manifest entries.
   - Reject a concurrent non-overwrite commit that would mix bucket counts in 
one partition.
   - Preserve the old `bucket-postpone` behavior when 
`postpone.batch-write-fixed-bucket=false`.
   - Update the Ray write documentation and regression test.
   
   Streaming writes are unchanged.
   
   ### Tests
   
   - `table_write_test.py`, `file_store_commit_test.py`, 
`table_commit_test.py`, and `conflict_detection_test.py`: 76 passed, 3 subtests 
passed.
   - Ray postpone-bucket round trip: 1 passed; all files landed in real bucket 
0 and all 10 rows were readable.
   - Python 3.6 `py_compile`: passed.
   - `git diff --check`: passed.


-- 
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