yiqijiu commented on PR #5556: URL: https://github.com/apache/hive/pull/5556#issuecomment-2509793221
> @yiqijiu Thanks for trying. Let's say I executed `mvn test -Pitests -Dtest=TestMiniLlapLocalCliDriver -Dtest.output.overwrite=true -Dqfile=okumin_new_test.q` with the set of your queries. In this case, the integration test system would generate `ql/src/test/results/clientpositive/okumin_new_test.q.out`, dumping the result of all queries including EXPLAIN. Bacause of the test configuration[1], we can see both # of rows and data size in `okumin_new_test.q.out`. With the master branch, I observed a negative data size(-8773271517492755968) in `okumin_new_test.q.out`. > > ``` > + Map 8 > + Map Operator Tree: > + TableScan > + alias: explain_multiple_ptf_big_table2 > + Statistics: Num rows: 9223372036854775 Data size: 4836736278108397824 Basic stats: COMPLETE Column stats: NONE > + Select Operator > ... > + Statistics: Num rows: 18446744073709550 Data size: -8773271517492755968 Basic stats: PARTIAL Column stats: NONE > ... > + Reducer 3 > + Execution mode: llap > + Reduce Operator Tree: > + Select Operator > + expressions: KEY.reducesinkkey0 (type: string), VALUE._col0 (type: string), KEY.reducesinkkey1 (type: string) > + outputColumnNames: _col0, _col1, _col2 > + Statistics: Num rows: 18446744073709550 Data size: -8773271517492755968 Basic stats: PARTIAL Column stats: NONE > ``` > > I expect all numbers will be positive with your nice patch. If we re-introduced the same mistake, the `.q.out` would fail, and a maintainer would notice the degradation. That's because Hive's CI checks the output of `*.q.out` is consistent with the previous revision. > > Please also make `numRows` more than `Long.MAX_VALUE / 2` so that we can test # of rows as well. > > - [1] If I remember correctly, [hive.explain.user=false](https://github.com/apache/hive/blob/rel/release-4.0.1/data/conf/llap/hive-site.xml#L288-L291) is the one that controls the behavior @okumin Yes, that’s exactly what I had in mind. Thank you for your patient guidance. I will try to write a qtest over the weekend, as I am quite busy during weekdays. Currently, my commute takes around 3.5 hours daily, so I only have a bit of time in the evenings, which is why I sometimes can’t respond to your help promptly. I may have some questions later on and will need to trouble you again -- 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: gitbox-unsubscr...@hive.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org For additional commands, e-mail: gitbox-h...@hive.apache.org
