Ungrouped aggregate function pushdown for duckdb#8645
Open
myrrc wants to merge 1 commit into
Open
Conversation
Signed-off-by: Mikhail Kot <mikhail@spiraldb.com>
myrrc
added
changelog/performance
6 tasks
6 tasks
Contributor
Author
|
As we don't have statistics attached to arrays, and we don't use precomputed statistics as Datafusion does, this PR would likely sit idle for some time until these are done, because currently there's no performance improvement compared to the default versions. There will be significant improvement once we add zone map statistics |
myrrc
changed the title
github-actions
Bot
removed
the
action/benchmark
Contributor
Polar Signals Profiling ResultsLatest Run
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals ProfilingVortex (geomean): 0.996x ➖ How to read Verdict and Engines
datafusion / vortex-file-compressed (0.996x ➖, 2↑ 1↓)
No file size changes detected. |
Contributor
Benchmarks: TPC-H SF=1 on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.022x ➖, 0↑ 1↓)
datafusion / vortex-compact (1.007x ➖, 0↑ 0↓)
datafusion / parquet (1.031x ➖, 0↑ 0↓)
datafusion / arrow (1.076x ➖, 1↑ 7↓)
duckdb / vortex-file-compressed (0.999x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.001x ➖, 0↑ 0↓)
duckdb / parquet (0.999x ➖, 2↑ 0↓)
duckdb / duckdb (0.995x ➖, 0↑ 0↓)
File Size Changes (10 files changed, -0.1% overall, 3↑ 7↓)
Totals:
|
Contributor
🚨🚨🚨❌❌❌ SQL BENCHMARK FAILED ❌❌❌🚨🚨🚨Benchmark |
Contributor
Benchmarks: FineWeb NVMeVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.014x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.030x ➖, 0↑ 1↓)
datafusion / parquet (1.008x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.996x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.986x ➖, 0↑ 0↓)
duckdb / parquet (1.016x ➖, 0↑ 0↓)
File Size Changes (1 files changed, +0.0% overall, 1↑ 0↓)
Totals:
|
Contributor
Benchmarks: Statistical and Population GeneticsVerdict: No clear signal (low confidence) How to read Verdict and Engines
duckdb / vortex-file-compressed (1.008x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.003x ➖, 0↑ 0↓)
duckdb / parquet (0.997x ➖, 0↑ 0↓)
File Size Changes (1 files changed, +0.0% overall, 1↑ 0↓)
Totals:
|
Contributor
Benchmarks: FineWeb S3Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.076x ➖, 1↑ 1↓)
datafusion / vortex-compact (1.144x ➖, 0↑ 3↓)
datafusion / parquet (1.124x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (1.014x ➖, 0↑ 0↓)
duckdb / vortex-compact (1.070x ➖, 0↑ 2↓)
duckdb / parquet (1.000x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Clickbench Sorted on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.073x ➖, 0↑ 3↓)
datafusion / parquet (1.064x ➖, 0↑ 3↓)
duckdb / vortex-file-compressed (0.979x ➖, 2↑ 1↓)
duckdb / parquet (1.039x ➖, 0↑ 0↓)
duckdb / duckdb (1.021x ➖, 0↑ 0↓)
File Size Changes (201 files changed, -0.0% overall, 111↑ 90↓)
Totals:
|
Contributor
Benchmarks: TPC-H SF=10 on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.944x ➖, 0↑ 0↓)
datafusion / vortex-compact (0.962x ➖, 0↑ 0↓)
datafusion / parquet (0.949x ➖, 0↑ 0↓)
datafusion / arrow (0.924x ➖, 5↑ 0↓)
duckdb / vortex-file-compressed (0.957x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.945x ➖, 0↑ 0↓)
duckdb / parquet (0.974x ➖, 0↑ 0↓)
duckdb / duckdb (0.965x ➖, 0↑ 0↓)
File Size Changes (27 files changed, -0.0% overall, 10↑ 17↓)
Totals:
|
Merging this PR will not alter performance
Warning Please fix the performance issues or acknowledge them on CodSpeed. Performance Changes
Tip Investigate this regression by commenting Comparing 
Footnotes
|
Contributor
Benchmarks: Clickbench on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.051x ➖, 2↑ 7↓)
datafusion / parquet (1.045x ➖, 0↑ 5↓)
duckdb / vortex-file-compressed (1.032x ➖, 3↑ 3↓)
duckdb / parquet (1.022x ➖, 0↑ 1↓)
duckdb / duckdb (1.036x ➖, 0↑ 0↓)
File Size Changes (106 files changed, -0.0% overall, 50↑ 56↓)
Totals:
|
Contributor
Benchmarks: Random AccessVortex (geomean): 0.947x ➖ How to read Verdict and Engines
unknown / unknown (0.981x ➖, 2↑ 1↓)
|
Contributor
🚨🚨🚨❌❌❌ SQL BENCHMARK FAILED ❌❌❌🚨🚨🚨Benchmark |
Contributor
Benchmarks: TPC-H SF=1 on S3Verdict: No clear signal (environment too noisy confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (1.165x ➖, 0↑ 6↓)
datafusion / vortex-compact (1.061x ➖, 0↑ 3↓)
datafusion / parquet (1.167x ➖, 0↑ 6↓)
duckdb / vortex-file-compressed (0.935x ➖, 1↑ 0↓)
duckdb / vortex-compact (0.952x ➖, 0↑ 0↓)
duckdb / parquet (1.005x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: Appian on NVMEVerdict: No clear signal (low confidence) How to read Verdict and Engines
datafusion / vortex-file-compressed (0.863x ✅, 7↑ 0↓)
datafusion / parquet (0.882x ✅, 7↑ 0↓)
duckdb / vortex-file-compressed (0.896x ✅, 6↑ 0↓)
duckdb / parquet (0.891x ✅, 5↑ 0↓)
duckdb / duckdb (0.910x ➖, 5↑ 0↓)
File Size Changes (4 files changed, -0.0% overall, 1↑ 3↓)
Totals:
|
Contributor
Benchmarks: CompressionVortex (geomean): 0.996x ➖ How to read Verdict and Engines
unknown / unknown (0.965x ➖, 13↑ 2↓)
|
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
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.
This PR adds support for pushing down ungropped aggregates min,max,sum,avg,mean,count(col), and count(*).
It consists of roughly 3 parts:
I've tried the implementation with CountStart being a separate accumulator, and it is much more complex than current one, because you can't provide a real column to accumulate on in select(), and you need to create a dummy one. So, we pay the cost of a separate atomic and a count_star positions vector, but our expression projection handling stays simple.
We also reject count(*) if it's the only aggregate because duckdb calculating it natively (just using chunk lengths) is faster than our accumulator pipeline
On the API input note, nearly all aggregate functions except for count_star() in duckdb have one non-const argument so input is a pair of a column and an expression.