Remove codepath asymmetry in dataframe count() #9136
Merged
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
|
on a mixed type frame this will cause a big drop in perf as the .values will convert everything to object (and actually may not work correctly) - so I believe some tests should fail and need a more comprehensive perf metric the notnull(frame) will do a block by block comparison |
|
@jreback Test suite passes for me and here's a full vbench run: |
|
the isnull determination is prob ok, but the vbenches are all very misleading. try this. In fact should add a mixed-type vbench. |
|
@qwhelan so I think you can use this perf fix, but only if |
|
@jreback Thanks for the pointer. I'll add a mixed-type case to the vb_suite and investigate this further in the next few days (getting on a plane in a few hours). |
|
@qwhelan thanks! certainly can be perf improved. |
|
Still have some cleanup to do, but here's a new vbench with the mixed-dtype case: |
|
@jreback I think this patch is ready - let me know if there's anything I should address. The axis information doesn't seem to be relevant as to which branch to take here, as |
|
@qwhelan ok this looks gr8! ping when green! and feel free to look for more of these types of things. Anytime And of course not everything is profiled (though hopefully tested) |
|
@jreback Travis is giving this a green - looks good to go. I'm seeing some good candidates for investigation, so I'll try and find some time to dig in over the next few weeks. |
jreback
added a commit
that referenced
this pull request
Dec 24, 2014
Remove codepath asymmetry in dataframe count()
|
@qwhelan thank you sir! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
@jreback I noticed a codepath asymmetry in core.frame.count that leads to a substantial difference in dropna() performance depending on the axis. Using the path
df.dropna(axis=0)takes yields a 2.5-5x improvement.