Open
Description
we could use a better error message rather than AttributeError
e.g. #38804 (comment)
when using
df.rolling(...).apply(f, raw=True)
when f
is using non-numpy functions e.g. .iloc
we could use a better error message rather than AttributeError
e.g. #38804 (comment)
when using
df.rolling(...).apply(f, raw=True)
when f
is using non-numpy functions e.g. .iloc