Fix scalar partial ordering comparison#6999
Merged
Merged
Conversation
Signed-off-by: Connor Tsui <connor.tsui20@gmail.com>
Merging this PR will improve performance by 18.36%
Performance Changes
Comparing 
Footnotes
|
robert3005
previously requested changes
Mar 18, 2026
robert3005
left a comment
robert3005
left a comment
Contributor
There was a problem hiding this comment.
we are missing some imports here
connortsui20
force-pushed
the
ct/scalar-partial-ord
branch
from
0f6645b to
24a99fc
Compare
joseph-isaacs
added
the
action/benchmark
Contributor
|
lets run benchmarks first just to check for regressions |
github-actions
Bot
removed
the
action/benchmark
Contributor
Polar Signals Profiling ResultsLatest Run
Powered by Polar Signals Cloud |
Contributor
Benchmarks: PolarSignals ProfilingVortex (geomean): 1.026x ➖ datafusion / vortex-file-compressed (1.026x ➖, 0↑ 0↓)
|
Contributor
Benchmarks: TPC-H SF=1 on NVMEVerdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (1.009x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.008x ➖, 0↑ 0↓)
datafusion / parquet (1.028x ➖, 1↑ 2↓)
datafusion / arrow (1.048x ➖, 1↑ 2↓)
duckdb / vortex-file-compressed (1.018x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.022x ➖, 0↑ 0↓)
duckdb / parquet (1.018x ➖, 3↑ 3↓)
duckdb / duckdb (1.013x ➖, 1↑ 2↓)
Full attributed analysis
|
Contributor
Benchmarks: FineWeb NVMeVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (1.159x ❌, 0↑ 7↓)
datafusion / vortex-compact (1.037x ➖, 0↑ 0↓)
datafusion / parquet (1.087x ➖, 0↑ 4↓)
duckdb / vortex-file-compressed (1.129x ❌, 0↑ 6↓)
duckdb / vortex-compact (1.076x ➖, 0↑ 1↓)
duckdb / parquet (1.066x ➖, 0↑ 1↓)
Full attributed analysis
|
Contributor
Benchmarks: TPC-DS SF=1 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (1.000x ➖, 1↑ 0↓)
datafusion / vortex-compact (0.990x ➖, 2↑ 0↓)
datafusion / parquet (0.985x ➖, 2↑ 0↓)
duckdb / vortex-file-compressed (1.002x ➖, 0↑ 1↓)
duckdb / vortex-compact (0.999x ➖, 1↑ 2↓)
duckdb / parquet (0.997x ➖, 1↑ 0↓)
duckdb / duckdb (0.987x ➖, 6↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: TPC-H SF=10 on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.941x ➖, 7↑ 0↓)
datafusion / vortex-compact (0.901x ➖, 13↑ 0↓)
datafusion / parquet (0.947x ➖, 5↑ 0↓)
datafusion / arrow (0.911x ➖, 9↑ 0↓)
duckdb / vortex-file-compressed (0.887x ✅, 14↑ 0↓)
duckdb / vortex-compact (1.000x ➖, 0↑ 0↓)
duckdb / parquet (0.997x ➖, 0↑ 0↓)
duckdb / duckdb (0.953x ➖, 0↑ 0↓)
Full attributed analysis
|
Signed-off-by: Connor Tsui <connor.tsui20@gmail.com>
connortsui20
force-pushed
the
ct/scalar-partial-ord
branch
from
24a99fc to
d541313
Compare
Contributor
Benchmarks: FineWeb S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.907x ➖, 0↑ 0↓)
datafusion / vortex-compact (1.078x ➖, 0↑ 3↓)
datafusion / parquet (0.963x ➖, 0↑ 0↓)
duckdb / vortex-file-compressed (0.993x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.962x ➖, 0↑ 0↓)
duckdb / parquet (0.983x ➖, 0↑ 0↓)
Full attributed analysis
|
Member
Author
|
test failure fixed, but that shouldnt stop the benchmarks from running |
Contributor
Benchmarks: TPC-H SF=1 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (0.713x ➖, 8↑ 0↓)
datafusion / vortex-compact (1.029x ➖, 1↑ 3↓)
datafusion / parquet (1.083x ➖, 0↑ 6↓)
duckdb / vortex-file-compressed (0.899x ➖, 1↑ 0↓)
duckdb / vortex-compact (0.970x ➖, 0↑ 0↓)
duckdb / parquet (1.011x ➖, 0↑ 1↓)
Full attributed analysis
|
Contributor
Benchmarks: Random AccessVortex (geomean): 0.864x ✅ unknown / unknown (0.960x ➖, 10↑ 4↓)
|
Contributor
Benchmarks: Statistical and Population GeneticsVerdict: No clear signal (low confidence) duckdb / vortex-file-compressed (1.041x ➖, 0↑ 1↓)
duckdb / vortex-compact (1.004x ➖, 0↑ 0↓)
duckdb / parquet (1.008x ➖, 0↑ 0↓)
Full attributed analysis
|
Contributor
Benchmarks: Clickbench on NVMEVerdict: No clear signal (low confidence) datafusion / vortex-file-compressed (0.911x ➖, 23↑ 1↓)
datafusion / parquet (0.907x ➖, 20↑ 0↓)
duckdb / vortex-file-compressed (0.921x ➖, 14↑ 0↓)
duckdb / parquet (0.942x ➖, 4↑ 0↓)
duckdb / duckdb (0.978x ➖, 6↑ 2↓)
Full attributed analysis
|
Contributor
Benchmarks: TPC-H SF=10 on S3Verdict: No clear signal (environment too noisy confidence) datafusion / vortex-file-compressed (1.023x ➖, 0↑ 1↓)
datafusion / vortex-compact (0.811x ➖, 4↑ 0↓)
datafusion / parquet (1.021x ➖, 0↑ 1↓)
duckdb / vortex-file-compressed (0.957x ➖, 0↑ 0↓)
duckdb / vortex-compact (0.901x ➖, 1↑ 0↓)
duckdb / parquet (0.967x ➖, 0↑ 0↓)
Full attributed analysis
|
joseph-isaacs
approved these changes
Mar 18, 2026
Contributor
Benchmarks: CompressionVortex (geomean): 0.994x ➖ unknown / unknown (1.054x ➖, 6↑ 54↓)
|
dimitarvdimitrov
pushed a commit
that referenced
this pull request
Mar 20, 2026
## Summary When we compare `Scalar`s, we have to be careful that we do not use `<` or `>` if we do not know that the scalars have different types. If they have different types, the result will not panic or raise an error, it will just return `false` for `>`. This changes the `scalar_cmp` function inside the comparison execution to use the `partial_cmp` method directly, where we map the optional result to an error and raise it. ## Testing Adds 2 regression tests, the first one is the one I stumbled on and the second is a more targeted one ## Unresolved Questions - Why is this function public? - Are there any other places that we do scalar comparison with the built-in `>` operators? If so, we may need to fix them. --------- Signed-off-by: Connor Tsui <connor.tsui20@gmail.com> Co-authored-by: Joe Isaacs <joe.isaacs@live.co.uk>
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.
Summary
When we compare
Scalars, we have to be careful that we do not use<or>if we do not know that the scalars have different types. If they have different types, the result will not panic or raise an error, it will just returnfalsefor>.This changes the
scalar_cmpfunction inside the comparison execution to use thepartial_cmpmethod directly, where we map the optional result to an error and raise it.Testing
Adds 2 regression tests, the first one is the one I stumbled on and the second is a more targeted one
Unresolved Questions
>operators? If so, we may need to fix them.