perf: dual replace #1064
Merged
perf: dual replace #1064
+2
−6
Conversation
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
Signed-off-by: Henry Schreiner <[email protected]>
henryiii
force-pushed
the
henryiii/perf/doublerepl
branch
from
712338f to
a459db8
Compare
KRRT7
approved these changes
Jan 19, 2026
notatallshaw
approved these changes
Jan 19, 2026
This was referenced Jan 19, 2026
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Outdated suggestions cannot be applied.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.
On Python 3.12 and 3.13,
str.translateis really slow. This is less elegant, but much faster on those two Python versions, and a bit faster on the other Python versions.This fixes a performance regression for Python 3.12 and 3.13 (#1030 is faster for all other Python versions), and is faster on all current Python versions.
(I suspect that
.translatecould be made faster in the future, as logically it could be faster, I think a lot more work went into optimizing.replace, there's a specialized fast path for one char replace) Edit: actually, not so sure. Translate is processing every char, while replace can do a vectorized copy, then scan looking for a single character replacement (there's a fast path for one char replacements). For a small number of chars, this is going to be faster than a hash table lookup and one byte builder. Optimizing translate to complete might be tricky.CC @hugovk, not as important for CPython, where this is only marginally faster, but very nice for libraries that run on 3.12 and 3.13.
str.translateis really slow in those versions.Credit to @sirosen for coming up with this alternative, I was thinking about falling back to the old method on CPython 3.12 and 3.13. :)