BUG: Categorical repr can exceed options.display.width

[wesm]

I see in the tests things like test_categorical_repr_datetime which exceeds the 80-character display width. This doesn't seem right to me, but want to check whether I'm missing something.

xref #15380

[wesm]

pandas/tests/test_categorical.py is a bloodbath because of this, so I'm going to disable flake8 checks in that file until this issue can be addressed.

[jreback]

In [9]: pd.set_option('display.width',120)
yeh I don't think the actual raw Categorical was repr was setup to limit printing, with the exception of the number of categories via display.max_categories

c = pd.Categorical(pd.date_range('20130101 09:00:00',periods=10))

In [10]: c
Out[10]: 
[2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00, 2013-01-04 09:00:00, 2013-01-05 09:00:00, 2013-01-06 09:00:00, 2013-01-07 09:00:00, 2013-01-08 09:00:00, 2013-01-09 09:00:00, 2013-01-10 09:00:00]
Categories (10, datetime64[ns]): [2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00, 2013-01-04 09:00:00, ..., 2013-01-07 09:00:00,
                                  2013-01-08 09:00:00, 2013-01-09 09:00:00, 2013-01-10 09:00:00]

In [11]: pd.set_option('display.width',80)

In [12]: c
Out[12]: 
[2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00, 2013-01-04 09:00:00, 2013-01-05 09:00:00, 2013-01-06 09:00:00, 2013-01-07 09:00:00, 2013-01-08 09:00:00, 2013-01-09 09:00:00, 2013-01-10 09:00:00]
Categories (10, datetime64[ns]): [2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00,
                                  2013-01-04 09:00:00, ..., 2013-01-07 09:00:00, 2013-01-08 09:00:00,
                                  2013-01-09 09:00:00, 2013-01-10 09:00:00]

In [13]: pd.set_option('display.max_categories',4)

In [14]: c
Out[14]: 
[2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00, 2013-01-04 09:00:00, 2013-01-05 09:00:00, 2013-01-06 09:00:00, 2013-01-07 09:00:00, 2013-01-08 09:00:00, 2013-01-09 09:00:00, 2013-01-10 09:00:00]
Categories (10, datetime64[ns]): [2013-01-01 09:00:00, 2013-01-02 09:00:00, ..., 2013-01-09 09:00:00,
                                  2013-01-10 09:00:00]

The index printer had a bit more effort

In [15]: pd.CategoricalIndex(c)
Out[15]: 
CategoricalIndex(['2013-01-01 09:00:00', '2013-01-02 09:00:00',
                  '2013-01-03 09:00:00', '2013-01-04 09:00:00',
                  '2013-01-05 09:00:00', '2013-01-06 09:00:00',
                  '2013-01-07 09:00:00', '2013-01-08 09:00:00',
                  '2013-01-09 09:00:00', '2013-01-10 09:00:00'],
                 categories=[2013-01-01 09:00:00, 2013-01-02 09:00:00, 2013-01-03 09:00:00, 2013-01-04 09:00:00, ...], ordered=False, dtype='category')