parquet: optimize CachedArrayReader byte-array coalescing#9743
Open
ClSlaid wants to merge 1 commit intoapache:mainfrom
Open
parquet: optimize CachedArrayReader byte-array coalescing#9743ClSlaid wants to merge 1 commit intoapache:mainfrom
ClSlaid wants to merge 1 commit intoapache:mainfrom
Conversation
When CachedArrayReader builds output from multiple cached batches, the old path materialized filtered byte arrays and then concatenated them. Replace that path for Utf8/Binary arrays with a direct coalescer that builds offsets, values, and validity in one output array, while keeping the existing generic MutableArrayData path for other types. Add a dedicated CachedArrayReader benchmark and a sparse string regression test so this path is measured directly and covered independently of broader parquet reader benchmarks. Benchmark vs main: - cached_array_reader/utf8_sparse_cross_batch_4m_rows/consume_batch: 11.949 ms -> 4.153 ms (-65.2%) - arrow_reader_clickbench/sync/Q24 (same filter/projection as ClickBench Q26): 28.377 ms -> 28.443 ms (+0.2%, no measurable change) Signed-off-by: cl <cailue@apache.org>
Contributor
Author
|
@alamb I've tried to optimize with GPT 5.4, the improvement is not that obvious in the original test case you gave. So I let it wrote a new benchmark and optimized on it. However, I'm still not really confident about the current implementation, so please have a look. |
Contributor
|
@XiangpengHao can you help review this PR? |
Contributor
|
run benchmarks arrow_reader_clickbench |
|
🤖 Arrow criterion benchmark running (GKE) | trigger CPU Details (lscpu)Comparing issue-9060-cached-array-reader-byte-coalescer (5e78671) to d7d9ad3 (merge-base) diff File an issue against this benchmark runner |
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Dandandan
reviewed
Apr 22, 2026
| selected_row_count: usize, | ||
| ) -> ArrayRef { | ||
| match selected_batches[0].array.data_type() { | ||
| ArrowType::Utf8 => { |
Contributor
There was a problem hiding this comment.
Wouldn't it need utf8view to show up in the clickbench benchmarks?
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
4 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.
When CachedArrayReader builds output from multiple cached batches, the old path materialized filtered byte arrays and then concatenated them. Replace that path for Utf8/Binary arrays with a direct coalescer that builds offsets, values, and validity in one output array, while keeping the existing generic MutableArrayData path for other types.
Add a dedicated CachedArrayReader benchmark and a sparse string regression test so this path is measured directly and covered independently of broader parquet reader benchmarks.
Benchmark vs main:
Which issue does this PR close?
Rationale for this change
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?