Fix SortPreservingRepartition with no existing ordering. #7811
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
alamb
approved these changes
Oct 12, 2023
There was a problem hiding this comment.
Thank you very much @mustafasrepo -- this would have taken me quite a while to figure out.
I reviewed this PR carefully and verified it solved my problem downstream.
| // corresponding SortExecs together. Also, the inputs of these `SortExec`s | ||
| // are not necessarily the same to be able to remove them. | ||
| let expected_input = ["BoundedWindowAggExec: wdw=[count: Ok(Field { name: \"count\", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }), frame: WindowFrame { units: Range, start_bound: Preceding(NULL), end_bound: CurrentRow }], mode=[Sorted]", | ||
| let expected_input = [ |
There was a problem hiding this comment.
The diff made it hard to see -- but I am pretty sure this test did not change, just the formatting did. Is that correct?
There was a problem hiding this comment.
Yes, just inserted new line
| " SortExec: expr=[a@0 ASC,b@1 ASC]", | ||
| " CsvExec: file_groups={1 group: [[x]]}, projection=[a, b, c, d, e], has_header=false", | ||
| ]; | ||
| let expected_optimized = [ |
There was a problem hiding this comment.
Without the code in this PR, I verified this test failed. The plan showed the same pattern as described in #7794
"BoundedWindowAggExec: wdw=[count: Ok(Field { name: \"count\", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }), frame: WindowFrame { units: Range, start_bound: Preceding(NULL), end_bound: CurrentRow }], mode=[Sorted]",
" SortPreservingMergeExec: []",
" SortExec: expr=[a@0 ASC,b@1 ASC]",
" SortPreservingRepartitionExec: partitioning=RoundRobinBatch(10), input_partitions=10",
" RepartitionExec: partitioning=RoundRobinBatch(10), input_partitions=1",
" CsvExec: file_groups={1 group: [[x]]}, projection=[a, b, c, d, e], has_header=false",
|
I don't see any reason to delay merging this fix. Thanks again @mustafasrepo |
mustafasrepo
commented
Oct 13, 2023
| " SortExec: expr=[nullable_col@0 DESC NULLS LAST]", | ||
| " ParquetExec: file_groups={1 group: [[x]]}, projection=[nullable_col, non_nullable_col], output_ordering=[nullable_col@0 ASC]"]; | ||
| let expected_optimized = ["WindowAggExec: wdw=[count: Ok(Field { name: \"count\", data_type: Int64, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {} }), frame: WindowFrame { units: Range, start_bound: CurrentRow, end_bound: Following(NULL) }]", | ||
| let expected_optimized = [ |
There was a problem hiding this comment.
similar to above case
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Which issue does this PR close?
Closes #7794.
Rationale for this change
What changes are included in this PR?
In the issue,
EnforceSortingmoves up the sort that satisfies the window. In this case, operators that are below theSortExecdo not have ordering at their inputs. However, some of them (SortPreservingRepartitionExec) expects its input to be ordered. While removingSortExecdown the plan, we also replace operators with their corresponding variants that do not preserve order. With this change I believeSortPreservingRepartitionExecwill be replaced withRepartitionExecin the issue.Are these changes tested?
Yes, a new test is added to replicate problem in the issue.
Are there any user-facing changes?