DOC: Update sort_index docs #31898
DOC: Update sort_index docs #31898
Conversation
dsaxton
commented
Feb 11, 2020
dsaxton
commented
- closes MultiLevelIndex sort_index already does support ascending as a list of booleans #31880
| If not None, sort on values in specified index level(s). | ||
| ascending : bool, default True | ||
| Sort ascending vs. descending. | ||
| ascending : bool or list of bools, default True |
There was a problem hiding this comment.
Can you update the annotation for this function to reflect this?
There was a problem hiding this comment.
So originally I had given it a Union type but mypy complained because here the ascending argument seemingly has to be a singleton bool: https://github.com/pandas-dev/pandas/blob/master/pandas/core/frame.py#L5055. I guess this is unavoidable because it has no way of knowing that it shouldn't get here when ascending is a list?
There was a problem hiding this comment.
Wouldn't that conflict with the documentation updates?
There was a problem hiding this comment.
Hmm, I think it's okay since it "can" accept a list of bools, just conditional on the index being a MultiIndex?
There was a problem hiding this comment.
Not sure - looks like this works (probably shouldn't)
>>> pd.DataFrame([[1]]).sort_index(ascending=[True, False, True])
0
0 1
| If not None, sort on values in specified index level(s). | ||
| ascending : bool, default true | ||
| Sort ascending vs. descending. | ||
| ascending : bool or list of bools, default True |
There was a problem hiding this comment.
if list of bools is allowed, should that show up in the annotation?
There was a problem hiding this comment.
See the discussion with WillAyd above; mypy doesn't seem to like that because nargsort gets called when we don't have a MultiIndex and it expects a single bool
|
@WillAyd Good to merge or were there changes you'd like me to make? Not sure if there's a "nice" way to get around the mypy annoyance |
|
Yea would be nice to get the typing figured out, but OK for another day |
|
Thanks @dsaxton |
4 participants
