TYP: func argument to DataFrame.apply, DataFrame.applymap, core.apply.frame_apply #38493
Conversation
pandas/core/frame.py
Outdated
| @@ -7656,7 +7658,13 @@ def transform( | |||
| return result | |||
|
|
|||
| def apply( | |||
| self, func, axis: Axis = 0, raw: bool = False, result_type=None, args=(), **kwds | |||
| self, | |||
| func: ApplyFuncType, | |||
There was a problem hiding this comment.
Just to make it explicit, this annotation will show up as Union[np.ufunc, Union[Callable, str, List[Union[Callable, str]], Dict[Union[Hashable, NoneType], Union[Callable, str, List[Union[Callable, str]]]]]]
(kind of unreadable, IMO, but so which is not even complete yet, since the ufunc should be present in each nested case Union[Callable, str]. But isn't a ufunc a callable, so that can maybe be removed?)
There was a problem hiding this comment.
Going off the aggregation case we don't special case ufunc there so we probably shouldn't do it here either.
In that case I think ApplyFuncType is the same as AggFuncType so we can use the same alias in both cases
There was a problem hiding this comment.
yeah i agree here, no need to be different, let's use AggFuncType
pandas/core/frame.py
Outdated
| @@ -7656,7 +7658,13 @@ def transform( | |||
| return result | |||
|
|
|||
| def apply( | |||
| self, func, axis: Axis = 0, raw: bool = False, result_type=None, args=(), **kwds | |||
| self, | |||
| func: ApplyFuncType, | |||
There was a problem hiding this comment.
yeah i agree here, no need to be different, let's use AggFuncType
|
thanks @arw2019 |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffSpinning off from #38416, xref #38416 (comment)