github-actions[bot] commented on code in PR #65392:
URL: https://github.com/apache/doris/pull/65392#discussion_r3544127742
##########
regression-test/suites/inverted_index_p2/test_show_data.groovy:
##########
@@ -813,13 +833,8 @@ suite("test_show_data_with_compaction", "p2") {
assertTrue(another_with_index_size != "wait_timeout")
logger.info("with_index_size is {}, another_with_index_size is {}",
with_index_size, another_with_index_size)
- // Index compaction merges per-segment index files; for identical data
the total
- // on-disk size may differ slightly from the non-compacted layout
(merge/file overhead).
- // Compare within 10% tolerance instead of exact equality to avoid
flakiness, while
- // still catching gross index bloat or corruption.
- assertTrue(Math.abs(with_index_size - another_with_index_size)
- <= 0.1 * Math.max(with_index_size,
another_with_index_size),
- "index size mismatch beyond 10% tolerance:
with_index=${with_index_size}, without_index=${another_with_index_size}")
Review Comment:
This drops the actual comparison between the compaction-enabled and
compaction-disabled paths. The suite deliberately builds one table with
`inverted_index_compaction_enable=true` and another with it `false`, but after
this change both values are only checked against broad independent ranges. For
example, in cloud mode 55 KB and 70 KB would both pass even though the two
paths differ by more than 20%, so a regression in compaction size accounting
could slip through. Please keep the absolute sanity range if needed, but also
retain a relative comparison between `with_index_size` and
`another_with_index_size`.
--
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]
