JingsongLi opened a new pull request, #9030:
URL: https://github.com/apache/paimon/pull/9030

   ## What changes
   
   - carry the planning snapshot in Java and Python full-text scan plans
   - pin deletion-vector live-row filtering and raw fallback reads to that 
snapshot
   - keep direct `read(splits)` calls unpinned and preserve primary-key 
full-text behavior
   - add regression coverage for commits that happen between planning and 
reading
   
   ## Why
   
   Data-evolution full-text scans selected index and raw ranges from one 
snapshot, but the reader evaluated deletion vectors and uncovered rows against 
the latest table state. A commit between planning and reading could therefore 
remove rows that were valid in the plan or otherwise make indexed and raw 
results inconsistent.
   
   The plan snapshot is exposed through a default Java method and an optional 
Python constructor argument, so existing plan implementations and callers 
remain compatible.
   
   ## Checks
   
   - `mvn -pl paimon-core -Pfast-build -DwildcardSuites=none 
-Dtest=FullTextSearchBuilderTest,PrimaryKeyFullTextReadTest,PrimaryKeyFullTextScanTest,PrimaryKeyFullTextSearchTest
 test` (38 tests)
   - `python3 -m pytest pypaimon/tests/vector_search_filter_test.py 
pypaimon/tests/primary_key_index_definitions_test.py -q` (88 tests, 7 subtests)
   - `mvn -pl paimon-core -DskipTests compile`
   - `python3 -m py_compile pypaimon/table/source/full_text_scan.py 
pypaimon/table/source/full_text_read.py 
pypaimon/tests/vector_search_filter_test.py`
   


-- 
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