jianguotian opened a new pull request, #11261:
URL: https://github.com/apache/paimon/pull/11261

   ### Purpose
   
   The pure Python planner can drop matching rows from fixed-bucket tables that 
use a custom Java bucket function (#10662). When the table sets 
`bucket-function.provider` and leaves `bucket-function.type` unset, the planner 
assumes the default hash. A point lookup can then skip the bucket that actually 
holds the key.
   
   Skip predicate-driven bucket pruning when the table declares a provider. 
Partition and file statistics remain available to narrow the scan.
   
   ### Tests
   
   The regression uses a four-bucket append table and explicitly places IDs 1 
and 2 in bucket 3, following the routing rule of the Java `test-last-bucket` 
provider. On master at `79b8ec4788`, the Python planner returns no rows for `id 
= 1`. With the change it returns exactly ID 1.
   
   - `PYPAIMON_TEST_EXECUTION=python python -m pytest -q 
pypaimon/tests/pushdown_bucket_test.py 
pypaimon/tests/read_builder_explain_test.py` passed all 58 tests.
   - Flake8 passed on both changed files with `dev/cfg.ini`.


-- 
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