Gabriel39 opened a new pull request, #68165:
URL: https://github.com/apache/doris/pull/68165
### What problem does this PR solve?
The Parquet join runtime filter case fails when a profile contains repeated
scan sections: `[98, 98, 3, 3]` violates the expected match count even though
the filtering threshold is satisfied.
Wait for profile completion through `ProfileAction`, extract
`RowGroupsFiltered` only from the fact table scan, and compare the largest
merged or instance count against the existing thresholds. Repeated sections are
not summed, and missing counters or insufficient filtering still fail. Only the
regression case changes.
### Release note
None
### Check List (For Author)
- Test
- [ ] Regression test
- [ ] Unit Test
- [x] Manual test
Local Groovy validation reproduced the original exact-count failure, then
checked normal, repeated, merged, multiple-instance, empty, missing-counter,
unrelated-table and insufficient-filtering inputs. Archived profile parsing
also passed with repeated sections and normalized whitespace. Framework polling
method checks covered delayed completion, transient read errors and timeouts.
The complete suite DSL executed with simulated SQL/profile responses for all
four queries and three wait settings; repeated insufficient counts were
rejected. `git diff --check` passed. Live Hive regression was not run.
- Behavior changed:
- [x] No production behavior changes; test assertions tolerate profile
layout variability.
- [ ] Yes.
- Does this need documentation?
- [x] No.
- [ ] Yes.
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label
--
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]