REGR: Allow positional arguments in DataFrame.agg

Although `DataFrame.agg` is documented as accepting `func, axis, *args, **kwargs`
with `*args` and `**kwargs` passed to `func`, passing positional
arguments raises a TypeError.

The reason for this is that the internal call to `self._aggregate` uses
a keyword argument (axis) before passing *args and `**kwargs`, and as such the
first positional argument is always interpreted as a second
specification of `axis`, which raises TypeError.

Prior to commit 433c900, TypeErrors
were being suppressed, falling back to `self.apply`, which in v1.1.0
turned into an error.

This fixes issue GH-36948.

Author: pganssle

pandas/core/frame.py

@@ -7407,7 +7407,7 @@ def aggregate(self, func=None, axis=0, *args, **kwargs):
 
         result = None
         try:
-            result, how = self._aggregate(func, axis=axis, *args, **kwargs)
+            result, how = self._aggregate(func, axis, *args, **kwargs)
         except TypeError as err:
             exc = TypeError(
                 "DataFrame constructor called with "