JingsongLi opened a new pull request, #1428:
URL: https://github.com/apache/paimon-rust/pull/1428

   ### Purpose
   
   `CALL sys.create_global_index(..., partitions => 'dt=2026-10-10')` currently 
fails with `Argument partitions is unknown`, although the core builder already 
supports partition filtering. Accept and apply this optional argument to match 
Paimon's Flink SQL and Spark SQL procedures.
   
   ### Brief change log
   
   - Bind `partitions` and pass the existing typed partition predicate to the 
global index builder, reusing the parser shared with drop and compaction 
procedures.
   - Cover BTree and full-text builds with AND/OR filters, typed values, NULL 
partitions, incremental coverage, and no-op repeated or unmatched builds.
   - Verify malformed filters and filters on unpartitioned tables fail without 
writing files.
   
   ### Tests
   
   - `RUST_MIN_STACK=16777216 cargo test --locked -p paimon-datafusion --test 
procedures` — 75 passed. The stack size matches the repository's DataFusion CI 
configuration.
   - `cargo +1.98.0 fmt --all -- --check` — passed.
   - `git diff --check` — passed.
   
   The new regressions fail before the fix because `partitions` is rejected as 
an unknown argument.
   
   ### API and Format
   
   Adds optional `partitions` to the existing `CALL sys.create_global_index` 
procedure. Omitting it retains whole-table build behavior. Snapshot and index 
file formats are unchanged.
   
   ### Documentation
   
   Updated `docs/src/sql.md` with partition filter syntax, NULL handling and 
no-op behavior, and corrected the existing documentation for partition-scoped 
index drops.
   


-- 
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