shuke987 opened a new pull request, #65382: URL: https://github.com/apache/doris/pull/65382
## Proposed changes Follow up the existing `test_segcompaction_agg_keys` deflake from #61026. The current master case already avoids the old golden file by allowing two concrete full-row results for `col_0=47`. This patch keeps the same intent but removes the hard-coded full-row candidates. The case now checks stable invariants instead: - `col_0=47` returns exactly one aggregate row. - The row has the expected 50 columns. - The key is `47`. - All `REPLACE` value columns belong to the known fruit value domain in the ORC input. This avoids pinning the test to a specific duplicate-key `REPLACE` winner while still checking that the load and aggregation returned a valid row. ## Why The ORC input contains multiple rows with `col_0=47`. With `AGGREGATE KEY(col_0)` and `REPLACE` value columns, the case should not require one specific source row to win unless the data/model makes that winner deterministic. ## Validation Remote branch-4.1 regression environment: ```bash ssh [email protected] cd /deploy/teamcity_agent13/work/a0477ef061788419/case_output/doris export JAVA_HOME=/usr/lib/jvm/java-8-openjdk-amd64 export PATH=$JAVA_HOME/bin:$PATH ./run-regression-test.sh --run -d segcompaction_p2 -s test_segcompaction_agg_keys ``` Result: ```text Test 1 suites, failed 0 suites, fatal 0 scripts, skipped 0 scripts ``` -- 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]
