brijrajk commented on PR #12151: URL: https://github.com/apache/gluten/pull/12151#issuecomment-4824739171
Force-pushed an update that **regenerates the TPC-DS plan-stability golden files** for `gluten-ut/spark40` and `gluten-ut/spark41`, and **drops** the earlier `[UT][VL] Refresh TPC-H q19 plan stability golden file` commit. **Why the goldens changed:** moving `BloomFilterMightContainJointRewriteRule` into the Operator Optimization batch (the logical-rule change from [this thread](https://github.com/apache/gluten/pull/12151#discussion_r-zhztheplayer)) changes the optimized plans for the ~21 TPC-DS queries that use runtime bloom filters — the producer/consumer now appear as `velox_bloom_filter_agg` / `velox_might_contain`. The checked-in `gluten-approved-plans-*` goldens were stale against the new rule, so `spark-test-spark40` and `spark-test-spark41` were failing on the plan-stability suites. This regenerates the affected `explain.txt`/`simplified.txt`; operator structure is unchanged for non-bloom queries. **Why the q19 commit was dropped:** TPC-H `q19` has no bloom filter, so its plan does not actually change under this PR. The earlier golden refresh only perturbed expression-id normalization (the `Brand#NN` string constants collide with ExprId numbers when a suite is regenerated in isolation), which did not match what CI generates. Dropping it leaves `q19` identical to `main` — which CI already passes on the spark41 side. **How they were regenerated (CI-matching):** all 7 plan-stability suites run together in one JVM with `SPARK_GENERATE_GOLDEN_FILES=1 SPARK_ANSI_SQL_MODE=false`, against Spark 4.0.2 and 4.1.1 respectively. (Running the suites together avoids the q19 ExprId-collision instability noted in `GlutenPlanStabilitySuite`.) This should clear the remaining `spark-test-spark40` / `spark-test-spark41` failures. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
