JingsongLi opened a new pull request, #10914:
URL: https://github.com/apache/paimon/pull/10914

   ### Purpose
   
   Primary-key indexes use separate `pk-*` properties, so the same global-index 
configuration cannot currently be reused across primary-key and Data Evolution 
tables. Resolve declarations and field options through shared `CoreOptions` 
entry points in Java and Python, while retaining each table mode's validation 
and index lifecycle.
   
   Support `global-index.<type>.index-columns`, the singular `index-column` 
alias, implementation-level parameters such as `ivf-pq.dimension`, and short 
keys in `fields.<column>.index-options`. Retain legacy PK declarations and 
field options, along with DLF flat vector field parameters.
   
   Apply resolved options throughout index construction, compaction, query 
planning, indexed and raw reads, vector refinement, and Spark local/distributed 
execution. Adapt native read/write schema transport and hybrid routes to the 
existing SDK contracts; qualify unsupported field-option combinations before 
execution. Update the documentation and add regression tests for compatibility 
and parameter propagation.
   
   ### Tests
   
   - Java Core: 252 tests passed, including option compatibility, schema 
validation, PK maintenance/reads, DE searches, composite BTree planning, and 
sorted writer rotation.
   - Spark (`-Pspark3`): 10 tests passed, covering local/distributed vector 
reads and sorted topology planning.
   - Flink (`-Pflink1`): 11 tests passed, covering sorted topology planning and 
filtered DE vector queries.
   - Python: 429 tests and 139 subtests passed; 34 tests skipped.
   - Maven validation ran without `fast-build`; formatting/checkstyle checks 
and `git diff --check` passed.
   
   The native runtime was unavailable locally, so native end-to-end execution 
remains unverified. Adapter boundary tests cover build options, read/write 
schema transport, query options, and hybrid route ABI constraints.
   


-- 
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