github-actions[bot] commented on code in PR #65672:
URL: https://github.com/apache/doris/pull/65672#discussion_r3588224466
##########
regression-test/suites/compaction/test_compaction_uniq_keys_with_delete.groovy:
##########
@@ -64,6 +64,7 @@ suite("test_compaction_uniq_keys_with_delete") {
`max_dwell_time` INT DEFAULT "0" COMMENT "用户最大停留时间",
`min_dwell_time` INT DEFAULT "99999" COMMENT "用户最小停留时间")
UNIQUE KEY(`user_id`, `date`, `datev2`, `datetimev2_1`,
`datetimev2_2`, `city`, `age`, `sex`) DISTRIBUTED BY HASH(`user_id`)
+ BUCKETS 1
Review Comment:
This still lets the test pass after only a partial cumulative compaction.
The loop below sums `rowset.split(" ")[1]` from `/api/compaction/show`, but BE
formats each rowset as `[start-end] num_segments DATA/DELETE ...`, so this is a
segment count. With this PR's one-bucket, one-replica layout the threshold is
`< 8`; compacting just two one-segment prefix rowsets into one output segment
can drop the total to 7 and exit even though later visible rowsets have not
been compacted. The final query only repeats the same logical result already
checked before compaction, so it does not prove the intended fully merged
state. Please make the wait condition check the actual final rowset/version
shape (or another exact compacted-state signal) in both changed tests instead
of relying on `< 8 * replicaNum`.
--
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]
