github-actions[bot] commented on code in PR #67882:
URL: https://github.com/apache/doris/pull/67882#discussion_r4002060932
##########
fe/fe-core/src/test/java/org/apache/doris/nereids/rules/exploration/mv/PartitionCompensatorTest.java:
##########
@@ -487,6 +487,84 @@ public void
testCalcInvalidPartitionsNoConcurrentModificationWithTwoRelatedTable
.forEach(v -> Assertions.assertEquals(expectedUnion, v));
}
+ @SuppressWarnings("unchecked")
+ @Test
+ public void
testCalcInvalidPartitionsCompensatesWholeIntersectingRollupBucket()
Review Comment:
[P1] Exercise the partial-bucket fix through the real rewrite path
This test hard-codes both the partition mapping and rewritten MV scan, then
stops at `calcInvalidPartitions`, so it proves that the helper now returns
`{p1, p2}` but not that the wrong-result path is fixed. It never exercises
production mapping expansion/freshness, `PartitionRemover`, `PredicateAdder`,
pruning invalidation and re-pruning, `LogicalUnion`, or result equality. The
cited unchanged `mtmv_range_date_part_up_rewrite` suite uses whole-table
queries and only stales disjoint months, where the old intersection and this
implementation behave identically. Please add a regression that makes one
multi-base-partition roll-up bucket stale, queries a proper subset of that
bucket plus a valid bucket, requires the MV union rewrite, and compares results
with rewrite disabled; keeping a disjoint stale bucket would also prove the
relevance guard.
--
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]
