[SPARK-39867][SQL][3.1] Global limit should not inherit OrderPreservingUnaryNode #37398
Conversation
This file contains 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
dongjoon-hyun
requested changes
Aug 4, 2022
…ryNode Make GlobalLimit inherit UnaryNode rather than OrderPreservingUnaryNode Global limit can not promise the output ordering is same with child, it actually depend on the certain physical plan. For all physical plan with gobal limits: - CollectLimitExec: it does not promise output ordering - GlobalLimitExec: it required all tuples so it can assume the child is shuffle or child is single partition. Then it can use output ordering of child - TakeOrderedAndProjectExec: it do sort inside it's implementation This bug get worse since we pull out v1 write require ordering. yes, bug fix fix test and add test Closes apache#37284 from ulysses-you/sort. Authored-by: ulysses-you <ulyssesyou18@gmail.com> Signed-off-by: Wenchen Fan <wenchen@databricks.com>
ulysses-you
force-pushed
the
SPARK-39867-3.1
branch
from
86b7e77
to
8669b9a
Compare
|
@dongjoon-hyun rebased |
viirya
approved these changes
Aug 5, 2022
|
The python syle check failure is unrelated. Thanks, merging to 3.1! |
cloud-fan
pushed a commit
that referenced
this pull request
Aug 5, 2022
…ngUnaryNode backport branch-3.1 for #37284 ### What changes were proposed in this pull request? Make GlobalLimit inherit UnaryNode rather than OrderPreservingUnaryNode ### Why are the changes needed? Global limit can not promise the output ordering is same with child, it actually depend on the certain physical plan. For all physical plan with gobal limits: - CollectLimitExec: it does not promise output ordering - GlobalLimitExec: it required all tuples so it can assume the child is shuffle or child is single partition. Then it can use output ordering of child - TakeOrderedAndProjectExec: it do sort inside it's implementation This bug get worse since we pull out v1 write require ordering. ### Does this PR introduce _any_ user-facing change? yes, bug fix ### How was this patch tested? fix test and add test Closes #37398 from ulysses-you/SPARK-39867-3.1. Authored-by: ulysses-you <ulyssesyou18@gmail.com> Signed-off-by: Wenchen Fan <wenchen@databricks.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
backport branch-3.1 for #37284
What changes were proposed in this pull request?
Make GlobalLimit inherit UnaryNode rather than OrderPreservingUnaryNode
Why are the changes needed?
Global limit can not promise the output ordering is same with child, it actually depend on the certain physical plan.
For all physical plan with gobal limits:
This bug get worse since we pull out v1 write require ordering.
Does this PR introduce any user-facing change?
yes, bug fix
How was this patch tested?
fix test and add test