PERF: periodarr_to_dt64arr #35171
Merged
PERF: periodarr_to_dt64arr #35171
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
jreback
requested changes
Jul 8, 2020
Contributor
jreback
left a comment
jreback
left a comment
There was a problem hiding this comment.
can you comment on the slowdowns on the particular cases, e.g. 4000, 5000, I guess we don't care about these as much
pandas/_libs/tslibs/period.pyx
Outdated
| Py_ssize_t i, l | ||
|
|
||
| l = len(periodarr) | ||
| if freq >= FR_DAY: |
Contributor
There was a problem hiding this comment.
can you add a comment here that you are short-circuiting for perf
pandas/_libs/tslibs/period.pyx
Outdated
| return ensure_datetime64ns(dta) | ||
|
|
||
| out = np.empty(l, dtype='i8') | ||
| l = len(periodarr) |
Contributor
There was a problem hiding this comment.
can you add a comment on which cases this hits
jreback
added
Performance
jreback
reviewed
Jul 8, 2020
pandas/_libs/tslibs/period.pyx
Outdated
|
|
||
| out = np.empty(l, dtype='i8') | ||
| l = len(periodarr) | ||
| out = np.empty(l, dtype="i8") |
Contributor
There was a problem hiding this comment.
this could be a nogil loop right? (actually couldn't this entire function)?
Member
Author
There was a problem hiding this comment.
this cant be nogil becuase it calls check_dts_bounds, which may raise
Member
Author
As mentioned in the OP, i dont understand why we see slowdowns this big. It should just be one extra |
…rf-periodarr_to_dt64arr
Member
Author
|
i think i might be able to fastpath some more cases, will give that a shot |
Member
Author
|
OK, i think i got these addressed: We still have some cases that see a 2-3x slowdown, but those are now all length-0 or length-1 cases where the actual time difference is tiny (and makes sense) |
jreback
approved these changes
Jul 9, 2020
Contributor
|
thanks |
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.
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffRe-use
ensure_datetime64nsfor the subset of cases where that is applicable. I'm at a loss for why we get such a slowdown for the other cases.