Optimize metric label cloning#22406
Open
xudong963 wants to merge 3 commits into
Open
Conversation
github-actions
Bot
added
physical-expr
Member
Author
|
run benchmarks |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (e049019) to c8b784a (merge-base) diff using: tpcds File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (e049019) to c8b784a (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (e049019) to c8b784a (merge-base) diff using: tpch File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagetpch — base (merge-base)
tpch — branch
File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagetpcds — base (merge-base)
tpcds — branch
File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
Member
Author
It shows similar performance improvements as lazy registration. |
alamb
approved these changes
May 21, 2026
Contributor
alamb
left a comment
alamb
left a comment
There was a problem hiding this comment.
I love it -- this is great @xudong963 🙏
I left some suggestions but I think this PR could be merged as is.
It seems like there may be a few more places where labels are created (like this) that we could (as a follow on PR) use the same trick:
datafusion/datafusion/physical-plan/src/repartition/mod.rs
Lines 1076 to 1080 in dc80bd7
| ) -> Self { | ||
| let name = name.into(); | ||
| let value = value.into(); | ||
| let name = LabelValue::from(name.into()); |
Contributor
There was a problem hiding this comment.
this is quite a nice way to avoid any public API changes 👍
| fn from(value: Cow<'static, str>) -> Self { | ||
| match value { | ||
| Cow::Borrowed(value) => Self::Static(value), | ||
| Cow::Owned(value) => Self::Shared(Arc::from(value)), |
Contributor
There was a problem hiding this comment.
This line is my only concern -- it requires an extra allocation (for the Arc) and deallocation (for the String) now.
We can probably remove that overhead as a follow on PR eventually.
Member
Author
There was a problem hiding this comment.
OKay, I'll try in a seperate PR
Comment on lines
535
to
537
| @@ -533,19 +536,19 @@ impl Label { | |||
| name: impl Into<Cow<'static, str>>, | |||
| value: impl Into<Cow<'static, str>>, | |||
Contributor
There was a problem hiding this comment.
We could potentially avoid the extra Arc allocation below by changing this sigature to
pub fn new(
name: impl Into<LabelValue>,
value: impl Into<LabelValue>,And then implement From<'static str> for LabelValue, From<Arc<str>> for LabelValue and From<String> for LabelValue and From<Cow<...>> for LabelValue
🤔
That way the API is the same, but you could create the Arc directly at the callsite
Contributor
|
FYI @Dandandan -- this is a great find / way to improve the performance of a bunch of the already really fast clickbench queries |
Contributor
|
run benchmark clickbench_partitioned |
46 tasks
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (e049019) to c8b784a (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
xudong963
force-pushed
the
optimize-metric-label-clone
branch
from
e049019 to
3a4b942
Compare
xudong963
force-pushed
the
optimize-metric-label-clone
branch
from
e344f5e to
fe340a8
Compare
Member
Author
|
run benchmark clickbench_partitioned |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (fe340a8) to c8b784a (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
Member
Author
|
run benchmark clickbench_partitioned |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing optimize-metric-label-clone (fe340a8) to c8b784a (merge-base) diff using: clickbench_partitioned File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usageclickbench_partitioned — base (merge-base)
clickbench_partitioned — branch
File an issue against this benchmark runner |
Contributor
|
One thing I find interesting is that improvement of delaying metric creation you did in #22353 seems to have a bigger positive improvement: As in there is more overhead to these metrics than just the string allocation it seems 🤔 |
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
3 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
ParquetFileMetrics::newregisters many per-file metrics with the samefilenamelabel. Before this PR, each metric built its own owned filename label withfilename.to_string(), which repeatedly copied the same dynamicstring during parquet scan setup.
This PR keeps parquet metrics eagerly registered, so
ExecutionPlan::metrics()visibility during execution is unchanged, while reducing repeated label string allocation and copying.What changes are included in this PR?
Labelname/value strings behindArc<str>internally, whilekeeping borrowed static label strings allocation-free.
filenamelabel across the per-file parquet metrics inParquetFileMetrics::new.and display the same way.
Are these changes tested?
Yes.
Are there any user-facing changes?
No. Metric registration timing and displayed label values are unchanged.