Optimize splitPart scalar function to reduce allocation #17660
+266
−9
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
justahuman1
commented
Feb 7, 2026
| * TODO: Revisit if index should be one-based (both Presto and Postgres use one-based index, which starts with 1) | ||
| * @param input | ||
| * @param delimiter | ||
| * @param input the input String to be split into parts. |
Author
There was a problem hiding this comment.
copied verbatim from the the overloaded splitPart function below to keep in sync.
justahuman1
force-pushed
the
split-part-malloc-fix
branch
from
33336f1 to
5a2b4f1
Compare
justahuman1
commented
Feb 7, 2026
| {"org.apache.pinot.common.function", ".", 3, 3, "common", "null"}, | ||
| {"+++++", "+", 0, 100, "", ""}, | ||
| {"+++++", "+", 1, 100, "null", "null"}, | ||
| {"+++++org++apache++", "", 1, 100, "null", "null"}, |
Author
There was a problem hiding this comment.
These were missing cases I added to validate identical behavior to previous version
justahuman1
commented
Feb 7, 2026
| @@ -0,0 +1,160 @@ | |||
| /** | |||
| * Licensed to the Apache Software Foundation (ASF) under one | |||
Author
There was a problem hiding this comment.
I can remove this if required. Added it for my own validation of the perf
Updated the splitPart function to use direct string traversal instead of allocating a full String[] array, significantly reducing memory pressure in high-throughput query scenarios. Performance improvements: - Space complexity: O(n) → O(1) - Time complexity: O(n) for both implementations Added JMH benchmarks to demonstrate improvements and regressions. The primary regression is in the backward index case with a large index value. This is the uncommon case and maybe worth the tradeoff since the memory allocation in the common case is now significantly reduced.
justahuman1
force-pushed
the
split-part-malloc-fix
branch
from
5a2b4f1 to
24644cf
Compare
justahuman1
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Optimized the splitPart function to use direct string traversal instead of allocating a full String[] array, significantly reducing memory pressure in high-throughput query scenarios.
Performance improvements:
Added JMH benchmarks to demonstrate improvements and regressions. The primary regression is in the backward index case so I added
splitPartNegativeIdxSingleCharDelim(b4272e2) which increases complexity but makes the kernel much fasterAddresses #17362
JMH Results
Yeah, the common cases are more than 10x faster 🚀 . The original results showed a regression with
negative_index, hence b4272e2. However, this code may be harder to manage so that is the trade-off. The below results are from all commits in the PRAllocation via
-prof gc