We only use it in one place, and removing it there doesn't break anything. Moreover, it behaves surprisingly: ``` >>> pd.Index(['a', np.nan, 'b']).is_mixed() True >>> Index([0, 'a', 1, 'b', 2, 'c']).is_mixed() False ```