JingsongLi opened a new pull request, #8997: URL: https://github.com/apache/paimon/pull/8997
## What changed - Reduce `LookupJoinITCase` from 87 test invocations to 22 by removing cache, predicate, projection, partition, sequence, and overwrite matrices already covered by direct lookup tests. - Keep SQL planner and Flink runtime boundaries such as temporal joins, retry behavior, asynchronous partition swaps, planner validation, and Blob SQL wiring. - Add direct tests for `BlobAsDescriptorRow` conversion and `DynamicPartitionNumberLoader` selection and refresh behavior. - Merge overlapping asynchronous refresh scenarios and remove waits which occurred before refresh was triggered. ## Why `LookupJoinITCase` took about 1,861 seconds in the Flink 1.x common test job and dominated its critical path. Most cases exercised lookup implementation semantics through a full Flink job even though the same contracts can be tested deterministically in seconds. This keeps end-to-end coverage at the planner/runtime boundary while moving implementation-level behavior to focused direct tests. No production code is changed. ## Impact - Test invocations: 87 -> 22 - Final local `LookupJoinITCase` runtime: 427.718 seconds - Direct lookup suite: 70 tests in 15.2 seconds Reference CI run: https://github.com/apache/paimon/actions/runs/30786551815/job/91601063500 ## Validation - `mvn -pl paimon-flink/paimon-flink-common -Pfast-build -Pflink1,spark3 -DwildcardSuites=none -Dtest=BlobAsDescriptorRowTest,DynamicPartitionNumberLoaderTest,DynamicPartitionLevelLoaderTest,LookupTableTest,FileStoreLookupFunctionTest test` - `mvn -pl paimon-flink/paimon-flink-common -Pfast-build -Pflink1,spark3 -DwildcardSuites=none -Dtest=LookupJoinITCase test` - `mvn -pl paimon-flink/paimon-flink-common -Pflink1,spark3 -DskipTests checkstyle:check spotless:check` -- 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]
