CLN: removes cython implementation of groupby count #11013
Conversation
|
|
||
| df = DataFrame({ | ||
| '1st':np.random.choice(list(ascii_lowercase), n), | ||
| '2nd':np.random.randint(0, 5, n), |
There was a problem hiding this comment.
add dtype='int64' for the ints, otherwise these fail on windows (default for ndarray is int32 there)
There was a problem hiding this comment.
this should work with any integer type; these are not the values passed to cython method
There was a problem hiding this comment.
of course - but it will fail on windows because the expected is int32
There was a problem hiding this comment.
no, it should not
There was a problem hiding this comment.
you try it, and if it crashed post the stack trace here
There was a problem hiding this comment.
pls make the changes that I commented then I will
There was a problem hiding this comment.
I stand corrected. You are not comparing versus the original, so ok here.
behzadnouri
force-pushed
the
grby-count
branch
from
5018fcb to
d968aab
Compare
| @@ -3468,6 +3459,24 @@ def _apply_to_column_groupbys(self, func): | |||
| in self._iterate_column_groupbys()), | |||
| keys=self._selected_obj.columns, axis=1) | |||
|
|
|||
| def count(self): | |||
There was a problem hiding this comment.
this needs to take the axis kw to be back-compat. (I see that you changed a test to fix this, pls change back).
There was a problem hiding this comment.
Having an axis argument there is a bug when current master simply ignores the axis arg; groupby.py:L807
also, when doing a groupby, counting on axis=1 has no meaning because u may get a different values for each row regardless of keys being the same or different.
CLN: removes cython implementation of groupby count
|
ok thanks |
No description provided.