-
Notifications
You must be signed in to change notification settings - Fork 232
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BUG] TpchLike query 2 fails when AQE is enabled #275
Comments
This is the error that happens, sometimes. Working theory is that there is a race condition driven by the order that concurrent query stages finish.
|
By controlling the order that stages finish, I can now reproduce this and we get this error in the case of a plan with a broadcast exchange wrapping a broadcast query stage:
|
The issue is that there is a difference in nullability of a referenced field between the required child partitioning and the actual child partitioning ( The attribute being referenced here is the result of a |
It is likely a difference in our code vs the original code. We replace operators in the physical plan after most of the planning has completed. |
Thanks for the pointer @revans2. I tracked this down to some bugs in GpuFilterExec. I will write up a separate issue for that. |
Root cause: #525 |
Signed-off-by: spark-rapids automation <70000568+nvauto@users.noreply.github.com>
Describe the bug
TpchLike query 2 fails when AQE is enabled
Steps/Code to reproduce bug
TpchLike query 2 is currently not run when AQE is enabled. Simply remove the conditional code in this scala test that disables it, and modify the Python test to run with both AQE and non-AQE configs just like all the other tests.
Expected behavior
The test should execute without error and produce the correct row count.
Environment details (please complete the following information)
Running
mvn clean verify
on desktop (Ubuntu).Additional context
N/A
The text was updated successfully, but these errors were encountered: