DEPR: idxmin/idxmax with all-NA

[jbrockmendel]

• closes BUG: pd.Series idxmax raises ValueError instead of returning <NA> when all values are <NA> #51276 (Replace xxxx with the GitHub issue number)
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

As discussed in last week's dev call.

[mroeschke]

Thanks @jbrockmendel

[jkittner]

Am I understanding this correctly that I'm supposed to write this now each and every time I use idxmax?:

s = pd.Series([None])

if not s.isnull().all():
    m = s.idxmax()
else:
    m = np.nan

instead of just this with the originally promised behavior of "If the entire Series is NA, the result will be NA."

s = pd.Series([None])
m = s.idxmax()

this is annoyingly verbose. Or what is the correct way to approach this?

[nkonts]

this is annoyingly verbose. Or what is the correct way to approach this?

Not only is it verbose, I find it less intuitive especially considering the parameter skipna.

pd.Series([None]).idxmax(skipna=True) resulting in np.nan (and similarly for pd.DataFrame.idxmax) is (for me) the expected behavior.