Adding Constant Check for FilterExec #9649
Conversation
github-actions
bot
added
core
|
A problem here is that when dealing with c in ('a', 'b', 'c') the parser would translate it into a BinaryOp::Eq and we could not directly do add the left part to the constant. So I made it a special case. |
Lordworms
changed the title
| @@ -131,7 +131,7 @@ fn update_coalesce_ctx_children( | |||
| coalesce_context.data = if children.is_empty() { | |||
| // Plan has no children, it cannot be a `CoalescePartitionsExec`. | |||
| false | |||
| } else if is_coalesce_partitions(&coalesce_context.plan) { | |||
| } else if is_coalesce_partitions(&coalesce_context.plan) || coalesce_context.data { | |||
There was a problem hiding this comment.
I think none of the changes in this file is necessary for this PR. I retracted changes in this file. All tests still pass
Co-authored-by: Mustafa Akur <106137913+mustafasrepo@users.noreply.github.com>
| let data_vec: Vec<_> = adjusted.children.iter().map(|x| x.data).collect(); | ||
| let mut plan_with_coalesce_partitions = | ||
| PlanWithCorrespondingCoalescePartitions::new_default(adjusted.plan); | ||
| plan_with_coalesce_partitions.children = plan_with_coalesce_partitions | ||
| .children | ||
| .into_iter() | ||
| .enumerate() | ||
| .map(|(id, mut child)| { | ||
| child.data = data_vec[id]; | ||
| child | ||
| }) | ||
| .collect(); |
There was a problem hiding this comment.
I think these changes are still unnecessary for this PR. Other than this, this PR is ready to merge. Thanks @Lordworms for this Pr.
|
Don't know why the win test failed, is this due to CI/CD issues? |
This unrelated to this Pr. There is an issue with windows runner. See issue: 9653 |
|
Got it, thanks |
Which issue does this PR close?
Closes #9612
Rationale for this change
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?