BUG: Index doesn't take over the Categorical name

[sinhrks]

Series can take over the Categorical.name.

import pandas as pd
c = pd.Categorical([1, 2, 3], name='xxx')
pd.Series(c)
#0    1
#1    2
#2    3
# Name: xxx, dtype: category
# Categories (3, int64): [1, 2, 3]

But Index doesn't.

pd.Index(c)
# CategoricalIndex([1, 2, 3], categories=[1, 2, 3], ordered=False, dtype='category')
pd.CategoricalIndex(c)
# CategoricalIndex([1, 2, 3], categories=[1, 2, 3], ordered=False, dtype='category')

[shoyer]

Why does Categorical have a name attribute at all? This seems strange to me, given that it's not a feature of other dtypes (which are just numpy arrays). @JanSchulz ?

[jankatins]

As far as I can see, that's a quite old property of Categorical from the time before a cat could be used as a Block.

This commit seems to have introduced the property: ea1186d

From reading this commit and #292, it seems that the name is needed when a Categorical is (or was?) used in groupbys.

[jorisvandenbossche]

My first reaction is the same: categoricals shouldn't have a name at all. @JanSchulz any idea if this now still needed?

[jankatins]

@jorisvandenbossche See #10632 At least the changes all seem ok...

[sinhrks]

name is being deprecated in #10632.