support suffix white space in arrow-cast parse#10396
Merged
alamb merged 11 commits intoJul 25, 2026
Merged
Conversation
Contributor
|
run benchmark cast_kernels |
This comment has been minimized.
This comment has been minimized.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
|
🤔 it seems to be slower |
Contributor
|
run benchmark cast_kernels |
1 similar comment
Contributor
|
run benchmark cast_kernels |
This comment has been minimized.
This comment has been minimized.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
This comment has been minimized.
This comment has been minimized.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
Author
|
yea its consistently slower 🤔 for the case where there is no white space we do three comparissions (1 for the prefix, 1 for the suffix, 1 for the suffix again) |
Contributor
|
run benchmark cast_kernels |
This comment was marked as duplicate.
This comment was marked as duplicate.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
Author
|
I don't think theres a way to avoid the extra checks white-spase stripping requires. Even when there is no extra space, this pr adds a prefix check, a suffix check and another suffix check just to do nothing else. we could add a conditional compilation attribute but that may be overkill. |
Contributor
|
im just curious how the previous PR seemed to suggest 0 regressions, meanwhile adding one more causes a noticeable bump though we are at the single digit microsecond scale |
Contributor
Author
@Jefffrey i am as-well. I think it has to do with the number of checks. In the previous PR, we only added three comparisons:
this PR has those exact same checks but twice and we also have two exta comparissions for the so about 8 extra comparisons. we could add something like |
Contributor
Author
|
mabey? #[inline]
fn trim_pre_and_post_whitespace(string: &str) -> &str {
let bytes = string.as_bytes();
let prefix = bytes.first().is_some_and(|b| !b.is_ascii_digit());
let suffix = bytes.last().is_some_and(|b| !b.is_ascii_digit());
match (prefix, suffix) {
(false, false) => string,
(true, false) => {
cold_path();
string.trim_ascii_start()
}
(false, true) => {
cold_path();
string.trim_ascii_end()
}
(true, true) => {
cold_path();
string.trim_ascii()
}
}
}this looks a bit clunky to me |
Contributor
|
something silly we could do is always assume the best case and have the worst case (need to trim) as a fallback impl Parser for Float32Type {
fn parse(string: &str) -> Option<f32> {
if let Ok(n) = lexical_core::parse(string.as_bytes()) {
Some(n)
} else {
let string = string.trim_ascii();
lexical_core::parse(string.as_bytes()).ok()
}
}
}since the benchmarks are testing the casting of arrays themselves (and not just a microbenchmark of this parse function) i think it would be good to aim for a way to avoid regressions if possible 🤔 that being said, its worth noting our benchmark only checks f32, and tests arrays of fixed 512 length, so its not like the suite is comprehensive to the changes being made here |
Contributor
Author
makes sense to me. Could we run the benchmarks again |
Contributor
|
run benchmark cast_kernels |
This comment was marked as duplicate.
This comment was marked as duplicate.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
|
run benchmark cast_kernels |
This comment was marked as duplicate.
This comment was marked as duplicate.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Contributor
Author
nice |
Jefffrey
reviewed
Jul 24, 2026
| fn parse(mut string: &str) -> Option<Self::Native> { | ||
| if !string.as_bytes().last().is_some_and(|x| x.is_ascii_digit()) { | ||
| return None; | ||
| fn parse(string: &str) -> Option<Self::Native> { |
Contributor
There was a problem hiding this comment.
perhaps we can add one benchmark for i32 just so we have a reference
Merged
Jefffrey
pushed a commit
that referenced
this pull request
Jul 24, 2026
# Which issue does this PR close? <!-- We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. --> - none. # Rationale for this change see #10396 (comment) <!-- Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed. Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes. --> # What changes are included in this PR? adds benchmarks for str -> i32 <!-- There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR. --> # Are these changes tested? n/a <!-- We typically require tests for all PRs in order to: 1. Prevent the code from being accidentally broken by subsequent changes 2. Serve as another way to document the expected behavior of the code If tests are not included in your PR, please explain why (for example, are they covered by existing tests)? If this PR claims a performance improvement, please include evidence such as benchmark results. --> # Are there any user-facing changes? <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. If there are any breaking changes to public APIs, please call them out. --> no
Contributor
|
run benchmark cast_kernels |
This comment was marked as duplicate.
This comment was marked as duplicate.
|
🤖 Arrow criterion benchmark completed (GKE) | trigger Instance: CPU Details (lscpu)Details
Resource Usagebase (merge-base)
branch
File an issue against this benchmark runner |
Jefffrey
approved these changes
Jul 24, 2026
Contributor
|
Merge up to get fix for MIRI |
alamb
approved these changes
Jul 25, 2026
alamb
left a comment
alamb
left a comment
Contributor
There was a problem hiding this comment.
Thank you @Rich-T-kid and @Jefffrey
| if let Ok(raw_float) = lexical_core::parse(string.as_bytes()) { | ||
| return Some(f16::from_f32(raw_float)); | ||
| } | ||
| let string = trim_pre_and_post_whitespace(string); |
Contributor
There was a problem hiding this comment.
this is a good strategy -- try the common/ fast path first and then fallback to checking whitespace afterwards
Contributor
|
🚀 |
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
4 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
What changes are included in this PR?
this PR adds support for suffix white space when calling parse
Are these changes tested?
yes
Are there any user-facing changes?
yes, users will be able to parse strings that have white space prefixed/suffixed onto numbers/floats.