perf: optimize arrays_zip_inner in datafusion-comet-spark-expr#4922
Draft
andygrove wants to merge 1 commit into
Draft
perf: optimize arrays_zip_inner in datafusion-comet-spark-expr#4922andygrove wants to merge 1 commit into
andygrove wants to merge 1 commit into
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Which issue does this PR close?
N/A — autonomous exploratory PR.
Rationale for this change
Borrow list offsets/null buffers instead of materializing per-column Vecs, size the MutableArrayData builders exactly, and coalesce unpadded rows into bulk extends so equal-length rows cost one copy per column instead of one call per row.
What changes are included in this PR?
Borrow list offsets/null buffers instead of materializing per-column Vecs, size the MutableArrayData builders exactly, and coalesce unpadded rows into bulk extends so equal-length rows cost one copy per column instead of one call per row.
How are these changes tested?
Correctness: unit tests + seeded differential fuzz (bit-identical Arrow output vs
main).Benchmark (criterion):
Full criterion output: