Multiindex.fillna() throws NotImplementedError but docs indicate it should work

[robbuckley]

Code Sample, a copy-pastable example if possible

idx = pd.MultiIndex.from_arrays([['a','b'],[1,np.nan]])
idx.fillna()
idx.values

Problem description

Error trace:

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
<ipython-input-63-feda4f0ee830> in <module>()
      1 idx = pd.MultiIndex.from_arrays([['a','b'],[1,np.nan]])
----> 2 idx.fillna()

~\Anaconda2\envs\python3\lib\site-packages\pandas\core\indexes\multi.py in fillna(self, value, downcast)
    784     def fillna(self, value=None, downcast=None):
    785         # isnull is not implemented for MultiIndex
--> 786         raise NotImplementedError('isnull is not defined for MultiIndex')
    787 
    788     @Appender(_index_shared_docs['dropna'])

NotImplementedError: isnull is not defined for MultiIndex

docs seem to state it should work: https://pandas.pydata.org/pandas-docs/stable/generated/pandas.MultiIndex.fillna.html

I am running v20.3 (anaconda doesnt seem to have 21.0 yet). But i've checked code and docs from both v21.0 and master and it seems to be the same

Expected Output

array([('a', 1.0), ('b', 999)], dtype=object)

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.3.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 78 Stepping 3, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.20.3
pytest: 3.2.1
pip: 9.0.1
setuptools: 36.5.0.post20170921
Cython: 0.26.1
numpy: 1.13.3
scipy: 0.19.1
xarray: 0.9.6
IPython: 6.1.0
sphinx: 1.6.3
patsy: 0.4.1
dateutil: 2.6.1
pytz: 2017.2
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.1.0
openpyxl: 2.4.8
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: 1.0.2
lxml: 4.1.0
bs4: 4.6.0
html5lib: 0.999999999
sqlalchemy: 1.1.13
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
pandas_gbq: None
pandas_datareader: None

[jreback]

very odd that the doc is showing up at all. Its simply not defined (though it is in a super-class), so maybe bleeding thru somehow.

cc @TomAugspurger

[TomAugspurger]

It is defined, but it just raises NotImplementedError. The docstring is inherited though.

We could override the docstring saying it's not implemented, but that's just a short term fix. I'm going to spend some time on our docs over the next couple weeks, so I hope to fix it properly then.

[jorisvandenbossche]

Yes, this is completely expected.

@robbuckley a short term fix of updating the MultiIndex.fillna docstring is still welcome if you want to do a PR!

[ghasemnaddaf]

@jorisvandenbossche plz see #18209