Repository navigation
bench(parquet): add level encoding and nested writer cases - #12213
Merged
Merged
Conversation
Add a level_encoding bench that feeds LevelEncoder::put_with_observer one mini-batch at a time with a counting and histogram observer, over random, alternating, run-length, sparse and list-shaped levels. Add list_short (List<Int32> with 0..=9 elements and nulls) and map cases to arrow_writer.
Contributor
|
run benchmark level_encoding |
|
🤖 Benchmark starting (GKE) | trigger Target: level_encoding Sharding factor: 1 (1 workers). Comparing bench-level-encoding-12194 (045b62d) to b2b7d47 (merge-base) diff Run configurationrun benchmark level_encoding
shards: 1
Results will be posted when all workers finish. File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Target: level_encoding Comparing bench-level-encoding-12194 (045b62d) to b2b7d47 (merge-base) diff Run configurationrun benchmark level_encoding
shards: 1
Baseline build unavailable for shard(s) 1; changed-only measurements for those workers. DetailsPer-runner informationlevel_encoding — shard 1/1Node: gk3-benchmark-cluster-nap-gneavwlm-88b578e7-dmr5 Instance: c4a-highmem-16 (12 vCPU / 65 GiB) uname: BENCH_COMMAND: cargo bench --features=arrow,async,test_common,experimental,object_store --bench level_encodingCPU Details (lscpu)Resource Usagebranch
File an issue against this benchmark runner |
Rich-T-kid
self-requested a review
November 12, 2026 17:36
Rich-T-kid
approved these changes
Nov 12, 2026
Rich-T-kid
left a comment
Contributor
There was a problem hiding this comment.
thanks @JonSnow1807 LGTM left a comment
| // What the column writer does per mini-batch: count values at max level | ||
| // and update the page's level histogram | ||
| group.bench_function(&name, |b| { | ||
| let mut encoder = LevelEncoder::v2_streaming(max_level); |
Contributor
There was a problem hiding this comment.
is there any specific reason why were using v2_streaming here?
Contributor
Author
There was a problem hiding this comment.
No particular reason beyond the encoding work being the same: the column writer uses v1_streaming for PARQUET_1_0 (the default) and v2_streaming for PARQUET_2_0, and both feed the same RleEncoder through put_with_observer. The only difference is the 4-byte length prefix v1 writes when a page is flushed. Happy to switch to v1_streaming to match the default if you prefer.
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?
Rationale for this change
Adds benchmarks for definition/repetition level encoding and more nested-column writer cases, so the change proposed in #12194 can be measured against main.
What changes are included in this PR?
parquet/benches/level_encoding.rs: level encoding benchmarks (mixed, uniform and fragmented levels).parquet/benches/arrow_writer.rs: two nested cases,list_short(short lists with nulls) andmap.Are these changes tested?
Benchmarks only; they build with cargo bench on main.
Are there any user-facing changes?
No.
Code drafted by both human and AI, and reviewed by human.