Pandas ENH: Add support for *args in apply-like functions.

Feature Type

• [ ] Adding new functionality to pandas

• [X] Changing existing functionality in pandas

• [ ] Removing existing functionality in pandas

Problem Description

I think it would be nice if all the various apply-like functions had the same way to pass extra arguments for the applied function. The state now is as follows: Series/DataFrame.apply - can accept **kwargs, but to pass non-keyword argument there you pass a tuple args, instead of *args. DataFrame.applymap - can accept **kwargs, but there is no way to pass non-keyword arguments. Index/Series.map - no way to pass any additional arguments. GroupBy.apply - can pass arguments using *args, **kwargs. GroupBy/Series/DataFrame.transform - can pass arguments using *args, **kwargs.

I think that the *args, **kwargs way is the cleanest, so I propose to change the implementation of all other cases to conform to it.

I also think that passing extra arguments to Series/DataFrame.apply using the args tuple should be deprecated (but while you have both options they should be mutually exclusive).

Feature Description

The implementation seems easy enough/already exists in most cases. For example in the Series.apply case:

def apply(func, convert_dtype=True, args=(), *nargs, **kwargs):
    if args != () and nargs != ():
        raise ValueError("Can only pass extra arguments using args or nargs, not both")
    args = args if args != () else nargs
    return SeriesApply(self, func, convert_dtype, args, kwargs).apply()

Alternative Solutions

As I said, it seems that in many places this change is already implemented and just needs to be applied everywhere.

Additional Context

No response

Comment From: Pvlb1998

Take

Comment From: Pvlb1998

Are the functions mentioned on the description (like Series/DataFrame.apply, DataFrame.applymap...) the only places where we need to apply the code on the solution?

Also, is there any possibility to show some examples where I can test these functions? It's just to make sure the code is working properly after the modifications.

Comment From: mroeschke

@Pvlb1998 just a heads up, please wait for confirmation from members of the core team regarding support of this functionality before moving forward.

Comment From: Pvlb1998

@mroeschke Oh it's alright! Even if this function doesn't properly work, I would still like to work on it purely for learning (doesn't have to be aproved). But even after making some modifications I still don't know how to test it. If it's possible, could you provide some cases of test so I can at least check if there's anything wrong with my modifications?

Comment From: rhshadrach

This looks like a duplicate of https://github.com/pandas-dev/pandas/issues/40112

Comment From: phofl

Yep agreed, closing then