BUG: Multiindex.nunique raises NotImplementedError

[mizuy]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Code Sample, a copy-pastable example

pd.DataFrame([[1,2],[1,2]]).set_index([0,1]).index.nunique()

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
<ipython-input-80-09716b89b157> in <module>
----> 1 pd.DataFrame([[1,2],[1,2]]).set_index([0,1]).index.nunique()

/opt/conda/lib/python3.7/site-packages/pandas/core/base.py in nunique(self, dropna)
   1284         uniqs = self.unique()
   1285         n = len(uniqs)
-> 1286         if dropna and isna(uniqs).any():
   1287             n -= 1
   1288         return n

/opt/conda/lib/python3.7/site-packages/pandas/core/dtypes/missing.py in isna(obj)
    124     Name: 1, dtype: bool
    125     """
--> 126     return _isna(obj)
    127 
    128 

/opt/conda/lib/python3.7/site-packages/pandas/core/dtypes/missing.py in _isna_new(obj)
    136     # hack (for now) because MI registers as ndarray
    137     elif isinstance(obj, ABCMultiIndex):
--> 138         raise NotImplementedError("isna is not defined for MultiIndex")
    139     elif isinstance(obj, type):
    140         return False

NotImplementedError: isna is not defined for MultiIndex

exactly same error raised by pandas 0.25.3

Problem description

The method exists, but always throws NotImplementedError. I can use len(df.index.unique()) as a substitute. In the previous version of pandas, documentation have Multiindex.nunique entry (https://pandas.pydata.org/pandas-docs/version/0.23.4/generated/pandas.MultiIndex.unique.html), but it has disappeared now. So I think it might be possible that this is not a bug, but an intended exception.

Expected Output

should be 1

Output of pd.show_versions()

INSTALLED VERSIONS

commit : None
python : 3.7.6.final.0
python-bits : 64
OS : Linux
OS-release : 4.19.76-linuxkit
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : en_US.UTF-8
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.0.3
numpy : 1.18.1
pytz : 2020.1
dateutil : 2.8.1
pip : 20.1
setuptools : 46.1.3.post20200325
Cython : 0.29.17
pytest : 5.4.1
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : 1.2.8
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.2
IPython : 7.14.0
pandas_datareader: None
bs4 : 4.9.0
bottleneck : 1.3.2
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.2.1
numexpr : 2.7.1
odfpy : None
openpyxl : 3.0.3
pandas_gbq : None
pyarrow : 0.17.0
pytables : None
pytest : 5.4.1
pyxlsb : None
s3fs : None
scipy : 1.4.1
sqlalchemy : 1.3.16
tables : 3.6.1
tabulate : None
xarray : 0.15.1
xlrd : 1.2.0
xlwt : None
xlsxwriter : 1.2.8
numba : 0.48.0

[dsaxton]

To me this feels like a bug, or at the very least is probably not the intended behavior. I think defining isna on a MultiIndex so this doesn't raise could make sense (or if not skip the NA check for MultiIndex). It does beg the question what it means for a MultiIndex value to be NA: do all the values need to be NA, or at least one (if the latter then the adjustment within nunique becomes more complicated)?

[FRidh]

Proposed fix in #38558.

[jbrockmendel]

This works in master, could use a test

[migunasekera]

take

[noatamir]

take