Repository navigation
Optimize byte-array dictionary gathering in arrow-row - #12113
Merged
Merged
Conversation
github-actions
Bot
added
arrow-row
Contributor
|
run benchmark row_format |
|
🤖 Benchmark starting (GKE) | trigger Target: row_format Sharding factor: 1 (1 workers). Comparing perf-dictionary-gather-12053 (9a086f3) to 8473490 (merge-base) diff Run configurationrun benchmark row_format
env:
BENCH_FILTER: dictionary
shards: 1
Results will be posted when all workers finish. File an issue against this benchmark runner |
Contributor
|
run benchmark row_format |
|
🤖 Benchmark starting (GKE) | trigger Target: row_format Sharding factor: 1 (1 workers). Comparing perf-dictionary-gather-12053 (9a086f3) to 8473490 (merge-base) diff Run configurationrun benchmark row_format
env:
BENCH_FILTER: dictionary
shards: 1
Results will be posted when all workers finish. File an issue against this benchmark runner |
Rich-T-kid
self-requested a review
November 7, 2026 15:48
|
🤖 Benchmark completed (GKE) | trigger Target: row_format Comparing perf-dictionary-gather-12053 (9a086f3) to 8473490 (merge-base) diff Run configurationrun benchmark row_format
env:
BENCH_FILTER: dictionary
shards: 1
DetailsPer-runner informationrow_format — shard 1/1Node: gk3-benchmark-cluster-nap-11fep6rz-1e90e640-9m46 Instance: c4a-highmem-16 (12 vCPU / 65 GiB) uname: BENCH_COMMAND: cargo bench --features=arrow,async,test_common,experimental,object_store --bench row_formatCPU Details (lscpu)Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Target: row_format Comparing perf-dictionary-gather-12053 (9a086f3) to 8473490 (merge-base) diff Run configurationrun benchmark row_format
env:
BENCH_FILTER: dictionary
shards: 1
DetailsPer-runner informationrow_format — shard 1/1Node: gk3-benchmark-cluster-nap-11fep6rz-4c5b963b-gqsw Instance: c4a-highmem-16 (12 vCPU / 65 GiB) uname: BENCH_COMMAND: cargo bench --features=arrow,async,test_common,experimental,object_store --bench row_formatCPU Details (lscpu)Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Rich-T-kid
reviewed
Nov 7, 2026
| } | ||
|
|
||
| #[test] | ||
| fn test_gather_dictionary_byte_values() { |
Contributor
There was a problem hiding this comment.
I think we can simplify this test, all binary types go through gather_dictionary_byte_values so just testing the stringArray values is fine
Contributor
Author
There was a problem hiding this comment.
Thanks for reviewing this. I’ve simplified test_gather_dictionary_byte_values to use StringArray, since the string and binary variants share the same gather implementation. The test still covers sliced keys, null keys, and keys that reference null dictionary values.
| } | ||
| } | ||
|
|
||
| fn gather_dictionary_byte_values<K: ArrowDictionaryKeyType, T: ByteArrayType>( |
Contributor
There was a problem hiding this comment.
can we add a small doc comment here
Contributor
Author
There was a problem hiding this comment.
Added a short doc comment to gather_dictionary_byte_values. It explains how the helper gathers values in dictionary-key order, preserves nulls, and returns an error if the accumulated offsets overflow.
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?
Closes #12053.
Rationale for this change
The sparse-dictionary encoding path gathers one dictionary value per input row before encoding. The current gather performs a dynamic MutableArrayData extension for every key, which adds overhead for common string and binary dictionaries.
What changes are included in this PR?
Add typed gather implementations for Utf8, LargeUtf8, Binary, and LargeBinary dictionary values. These build the offsets, value buffer, and validity bitmap directly.
Other value types continue to use the generic MutableArrayData implementation. Offset growth is checked so that overflow continues to reach the existing fallback. The dictionary-gather thresholds are unchanged.
Are these changes tested?
I’ve simplified test_gather_dictionary_byte_values to use StringArray, since the string and binary variants share the same gather implementation. The test still covers sliced keys, null keys, and keys that reference null dictionary values.
All 110 arrow-row library tests pass. Workspace tests, workspace Clippy checks, formatting, and spelling checks also pass.
Performance was measured using the existing prepared 8,192-row string-dictionary benchmark on an Intel Xeon E5-2673 v4, Linux x86_64, with Rust 1.99.1. This benchmark cycles through contiguous key chunks.
Two paired Criterion runs used 50 samples, a one-second warmup, and a three-second measurement period. Execution order was reversed for the second pair. The table reports the average of the two runs’ mean estimates.
The 10,000-value control does not enter the dictionary-gather path and averaged 2.5% slower in these measurements. Individual runs varied by several percent.
The additional dictionary sizes were used through a separate local benchmark patch; they are not included in this implementation diff.
Are there any significant user-facing changes?
No public API or intended result changes.
AI assistance: AI tools assisted with implementation, tests, independent review, and validation. I reviewed the code, tests, and description before submission.