Categorical.from_codes warns if None is in categories

[ssanderson]

Born from a gitter conversation with @jorisvandenbossche and @jreback.

As of #10748, it's deprecated to have np.NaN as a category label. The deprecation warning also fires if None is a category, but it's unclear whether this is intended behavior, and the error message explicitly refers to NaN, which is confusing.

Code Sample, a copy-pastable example if possible

pd.Categorical.from_codes([0, 1, 2], [None, 'a', 'b'])
Setting NaNs in `categories` is deprecated and will be removed in a future version of pandas.

Expected Output

I wouldn't have expected a warning about NaN on an array that doesn't contain NaN.

output of pd.show_versions()

In [6]: pd.show_versions() ## INSTALLED VERSIONS

commit: None
python: 2.7.10.final.0
python-bits: 64
OS: Linux
OS-release: 4.2.0-16-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.18.1
nose: 1.3.7
pip: 8.1.2
setuptools: 24.0.2
Cython: 0.22.1
numpy: 1.11.1
scipy: 0.15.1
statsmodels: 0.6.1
xarray: None
IPython: 3.2.1
sphinx: 1.3.4
patsy: 0.4.0
dateutil: 2.4.2
pytz: 2016.4
blosc: 1.2.8
bottleneck: 1.0.0
tables: None
numexpr: 2.4.6
matplotlib: 1.4.3
openpyxl: None
xlrd: 0.9.4
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: 1.0.8
pymysql: None
psycopg2: 2.6.1 (dt dec pq3 ext lo64)
jinja2: 2.7.3
boto: 2.38.0
pandas_datareader: 0.2.1

[sinhrks]

Because pandas regards None as NaN, it is intended. This is applied to pd.NaT also.

pd.Categorical.from_codes([0, 1, 2], [pd.NaT, 'a', 'b'])
# Setting NaNs in `categories` is deprecated and will be removed in a future version of pandas.

# [NaT, a, b]
# Categories (3, object): [NaT, a, b]

Appreciated if you're willing to fix warning to be more appropriate.

[ssanderson]

@sinhrks can you clarify what you mean by "pandas regards None as NaN"? It treats both as "null" values for many purposes, but NaN is its own interesting beast which generally requires logic different from None because of the fact that NaN != NaN.

[sinhrks]

I meant both are treated as missing, see here.
http://pandas.pydata.org/pandas-docs/stable/missing_data.html

[jreback]

@gfyoung I think #15806 removes the need for this issue? (also add a test with None as a category as well.

[gfyoung]

@jreback: Sure does. Will add tests to confirm.

[jorisvandenbossche]

I think the same question of @ssanderson still stands, as the message still says "Categorial categories cannot be NaN", while he uses None, which makes it a bit confusing.

[jreback]

@jorisvandenbossche I'll change the wording to null values.