perf: Improve performance of split_part
#19570
Open
+400
−13
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
andygrove
changed the title
split_part
andygrove
commented
Dec 30, 2025
| .try_for_each(|((string, delimiter), n)| -> Result<(), DataFusionError> { | ||
| match (string, delimiter, n) { | ||
| (Some(string), Some(delimiter), Some(n)) => { | ||
| let split_string: Vec<&str> = string.split(delimiter).collect(); |
Member
Author
There was a problem hiding this comment.
This was allocating strings for all parts even if only some parts were needed
Contributor
|
46x faster 👍 |
comphead
approved these changes
Dec 30, 2025
Contributor
comphead
left a comment
comphead
left a comment
There was a problem hiding this comment.
Thanks @andygrove the early return makes much more sense than eagerly calculating all the parts
viirya
approved these changes
Dec 30, 2025
Jefffrey
approved these changes
Dec 31, 2025
martin-g
reviewed
Dec 31, 2025
| std::cmp::Ordering::Greater => { | ||
| // Positive index: use nth() to avoid collecting all parts | ||
| // This stops iteration as soon as we find the nth element | ||
| string.split(delimiter).nth((n - 1) as usize) |
Member
There was a problem hiding this comment.
Are 32-bit systems supported ?
n is Int64, so it is possible that this cast may lead to a truncation or even a crash in debug build
| std::cmp::Ordering::Less => { | ||
| // Negative index: use rsplit().nth() to efficiently get from the end | ||
| // rsplit iterates in reverse, so -1 means first from rsplit (index 0) | ||
| string.rsplit(delimiter).nth((-n - 1) as usize) |
Member
There was a problem hiding this comment.
another corner case: -n will fail for i64::MIN
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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?
Rationale for this change
I ran microbenchmarks comparing DataFusion with DuckDB for string functions (see apache/datafusion-benchmarks#26) and noticed that DF was very slow for
split_part.This PR fixes some obvious performance issues. Speedups are:
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?