[CALCITE-4765] Complex correlated EXISTS sub-queries used as scalar subqueries can return wrong results #4724
Conversation
|
@NobiGo If you have time, could you please check if it's working correctly now? Since I couldn't find an |
mihaibudiu
left a comment
mihaibudiu
left a comment
There was a problem hiding this comment.
The CI is unhappy for some reason too
xiedeyantu
force-pushed
the
CALCITE-4765
branch
from
053b4a9 to
01e4845
Compare
Fixed, thanks! |
|
This is the result compared with PGSQL. |
…ubqueries can return wrong results
xiedeyantu
force-pushed
the
CALCITE-4765
branch
from
01e4845 to
9810a36
Compare
|
I tried constructing |
Is this the right link? |
Sorry, I have updated the original SQL to the SQL in the current PR. |
|
| !} | ||
|
|
||
| !if (use_new_decorr) { | ||
| # TODO: TopDownGeneralDecorrelator should fix trimmer error |
There was a problem hiding this comment.
Is anyone working on fixing this?
Should we wait for the fix?
There was a problem hiding this comment.
This is an issue of compatibility between the new decorrelator and Trimmer(I think this is a problem we expected). I think we could create a Jira instance to document this problem. Waiting until this issue is fixed before merging this PR is also a good option.
There was a problem hiding this comment.
If no one wants to work on it, I will try to fix the problem.
There was a problem hiding this comment.
Start by creating a new issue and check with @NobiGo, he has been working on the trimmer recently.
There was a problem hiding this comment.
If an issue is created, please @ me.
|
|
||
| !ok | ||
|
|
||
| # [CALCITE-4765] Complex correlated EXISTS sub-queries used as scalar subqueries can return wrong results |
There was a problem hiding this comment.
why is blank the right file for this test?
There was a problem hiding this comment.
Since no emps table with a commission column was found in quidem test, I need to create one. Alternatively, we can use !use blank in other .iq files.
| !} | ||
|
|
||
| !if (use_new_decorr) { | ||
| # TODO: TopDownGeneralDecorrelator should fix trimmer error |
There was a problem hiding this comment.
If an issue is created, please @ me.
| SELECT e2.deptno FROM emps e2 where e1.commission = e2.commission) as table3 | ||
| where table3.deptno <> e1.deptno) | ||
| from emps e1; | ||
| !if (use_old_decorr) { |
There was a problem hiding this comment.
Which one is used when running this specific test case?
There was a problem hiding this comment.
Thank you so much for your prompt attention to this issue. It involves new Quidm tests, new TopDownGeneralDecorrelator, and new MARK JOINs. Please feel free to @ me if you'd like to discuss it. @silundong If you have time, you can also look into this issue, because you might understand it better. Thanks！
|
@NobiGo @mihaibudiu I have file a new jira CALCITE-7356 to describe it. |
3 participants
See CALCITE-4765
Only add tests.