groupby/quantile breaks

[daishi]

Code Sample, a copy-pastable example if possible

In [34]: pd.DataFrame({'x': [0, 1, 2], 'y': [0, 1, 2]}).groupby('x')['y'].quantile([0.5])                                                                                                                          
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-34-287422cf6b27> in <module>
----> 1 pd.DataFrame({'x': [0, 1, 2], 'y': [0, 1, 2]}).groupby('x')['y'].quantile([0.5])

~/.venv-3.7.2/p4p_backend-1ulE3AqG/lib/python3.7/site-packages/pandas/core/groupby/groupby.py in quantile(self, q, interpolation)
   1951             indices = np.concatenate(arrays)
   1952             assert len(indices) == len(result)
-> 1953             return result.take(indices)
   1954 
   1955     @Substitution(name="groupby")

~/.venv-3.7.2/p4p_backend-1ulE3AqG/lib/python3.7/site-packages/pandas/core/series.py in take(self, indices, axis, is_copy, **kwargs)
   4430 
   4431         indices = ensure_platform_int(indices)
-> 4432         new_index = self.index.take(indices)
   4433 
   4434         if is_categorical_dtype(self):

~/.venv-3.7.2/p4p_backend-1ulE3AqG/lib/python3.7/site-packages/pandas/core/indexes/multi.py in take(self, indices, axis, allow_fill, fill_value, **kwargs)
   2030             allow_fill=allow_fill,
   2031             fill_value=fill_value,
-> 2032             na_value=-1,
   2033         )
   2034         return MultiIndex(

~/.venv-3.7.2/p4p_backend-1ulE3AqG/lib/python3.7/site-packages/pandas/core/indexes/multi.py in _assert_take_fillable(self, values, indices, allow_fill, fill_value, na_value)
   2058                 taken = masked
   2059         else:
-> 2060             taken = [lab.take(indices) for lab in self.codes]
   2061         return taken
   2062 

~/.venv-3.7.2/p4p_backend-1ulE3AqG/lib/python3.7/site-packages/pandas/core/indexes/multi.py in <listcomp>(.0)
   2058                 taken = masked
   2059         else:
-> 2060             taken = [lab.take(indices) for lab in self.codes]
   2061         return taken
   2062 

IndexError: index 3 is out of bounds for size 3

Problem description

An exception is thrown above. This was not an issue in Pandas 0.24.2.

Expected Output

Output of pd.show_versions()

INSTALLED VERSIONS
------------------
commit           : None
python           : 3.7.2.final.0
python-bits      : 64
OS               : Linux
OS-release       : 5.0.0-27-generic
machine          : x86_64
processor        : x86_64
byteorder        : little
LC_ALL           : None
LANG             : en_US.UTF-8
LOCALE           : en_US.UTF-8

pandas           : 0.25.1
numpy            : 1.17.1
pytz             : 2019.2
dateutil         : 2.8.0
pip              : 19.2.3
setuptools       : 41.2.0
Cython           : None
pytest           : 5.1.2
hypothesis       : None
sphinx           : None
blosc            : None
feather          : None
xlsxwriter       : 1.2.0
lxml.etree       : 4.4.1
html5lib         : None
pymysql          : None
psycopg2         : 2.8.3 (dt dec pq3 ext lo64)
jinja2           : 2.10.1
IPython          : 7.5.0
pandas_datareader: None
bs4              : 4.8.0
bottleneck       : None
fastparquet      : None
gcsfs            : None
lxml.etree       : 4.4.1
matplotlib       : 3.1.1
numexpr          : None
odfpy            : None
openpyxl         : 2.6.3
pandas_gbq       : None
pyarrow          : None
pytables         : None
s3fs             : None
scipy            : 1.3.1
sqlalchemy       : 1.3.8
tables           : None
xarray           : None
xlrd             : 1.2.0
xlwt             : None
xlsxwriter       : 1.2.0

[WillAyd]

Can you try on master? I believe this has already been fixed

[josesho]

I think I'm seeing the same issue here: #28312.

Basically, if the GroupBy column has more than 2 categories, the quantile code breaks.

[TomAugspurger]

Fixed by #28113 I think. That's included in 0.25.2 which will be out in a few weeks, or you can use master.

[daishi]

Confirming fixed on master:

In [4]: import pandas as pd                                                                                                                                                                                        

In [5]: pd.DataFrame({'x': [0, 1, 2], 'y': [0, 1, 2]}).groupby('x')['y'].quantile([0.5])                                                                                                                           
Out[5]: 
x     
0  0.5    0.0
1  0.5    1.0
2  0.5    2.0
Name: y, dtype: float64

In [6]: pd.__version__                                                                                                                                                                                             
Out[6]: '0.25.0+296.g2d65e38f5'

FWIW as a side note it took me a little bit to figure out how to build and run off of master. Also, I didn't initially anticipate the build time required for the native libraries. Happy to help but maybe a link giving a quick outline of process when suggesting "try on master" would help reduce friction for casual bug reporters like myself?