JingsongLi opened a new pull request, #10741:
URL: https://github.com/apache/paimon/pull/10741

   ## Purpose
   
   Complete the execution boundary for primary-key index planning. Public APIs, 
configuration, metadata types and wire formats remain unchanged.
   
   ## Changes
   
   - Move `PkSortedBucketIndexState` and `PkSortedIndexGroup` into 
`execution/python/index/pksorted`.
   - Move the primary-key index source policy into `execution/python/index/pk`.
   - Update the Python scalar, vector and full-text scanners and their internal 
tests to import these implementations from the execution layer. Remove the 
empty former `index/pksorted` package.
   - Keep index definitions, source metadata, shared dynamic-bucket assignment 
and `IndexFileHandler` in the shared layer.
   - Document the boundary and add a fresh-process regression that rejects 
every `execution.python` import while exercising public metadata, scalar 
planning and Native primary-key vector/full-text scan/read delegation, 
including query failure propagation.
   
   The moved algorithms are unchanged; only internal import paths change. No 
compatibility forwarding modules or new public APIs are introduced.
   
   ## Verification
   
   - 55 existing primary-key sorted-index, definition, Java golden-fixture and 
execution-boundary tests passed.
   - The new Native import-isolation regression passed on both PyArrow 19 and 
PyArrow 16 (56 distinct cases including the existing group).
   - Python 3.6 grammar, including the subprocess script, and 
repository-configured Flake8 4 passed for all 11 changed Python files.
   - Package discovery includes all three new packages. Diff checks passed, and 
independent review found no actionable issues.
   
   The Native boundary test mocks the SDK/FFI surface while exercising the real 
public facades and Native adapters. No Native extension was rebuilt or 
installed.
   


-- 
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