BUG: rolling aggregate() with a list of functions along axis 1 raises ValueError

[mvashishtha]

Pandas version checks

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

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

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd
import numpy as np

df = pd.DataFrame([[1, 2], [3, 4]])
# rolling across axis=0 works fine:
print(df.rolling(window=2, axis=0, min_periods=1).aggregate([np.sum, np.mean]))
# This causes a ValueError
print(df.rolling(window=2, axis=1, min_periods=1).aggregate([np.sum, np.mean]))

Issue Description

I start with the dataframe:

   0  1
0  1  2
1  3  4

Rolling over axis=0 gives the rolling sum and mean for each column:

     0         1
   sum mean  sum mean
0  1.0  1.0  2.0  2.0
1  4.0  2.0  6.0  3.0

Rolling over axis=1 should give rolling sum and mean for each row, but instead raises a ValueError: No axis named 1 for object type Series:

Show stack trace

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
File ~/pandas/pandas/core/generic.py:535, in NDFrame._get_axis_number(cls, axis)
    534 try:
--> 535     return cls._AXIS_TO_AXIS_NUMBER[axis]
    536 except KeyError:

KeyError: 1

During handling of the above exception, another exception occurred:

ValueError                                Traceback (most recent call last)
Input In [20], in <module>
      4 df = pd.DataFrame([[1, 2], [3, 4]])
      5 # rolling across axis=0 works fine:
----> 6 df.rolling(window=2, axis=1, min_periods=1).aggregate([np.sum, np.mean])

File ~/pandas/pandas/core/window/rolling.py:1765, in Rolling.aggregate(self, func, *args, **kwargs)
   1727 @doc(
   1728     _shared_docs["aggregate"],
   1729     see_also=dedent(
   (...)
   1763 )
   1764 def aggregate(self, func, *args, **kwargs):
-> 1765     return super().aggregate(func, *args, **kwargs)

File ~/pandas/pandas/core/window/rolling.py:635, in BaseWindow.aggregate(self, func, *args, **kwargs)
    634 def aggregate(self, func, *args, **kwargs):
--> 635     result = ResamplerWindowApply(self, func, args=args, kwargs=kwargs).agg()
    636     if result is None:
    637         return self.apply(func, raw=False, args=args, kwargs=kwargs)

File ~/pandas/pandas/core/apply.py:171, in Apply.agg(self)
    168     return self.agg_dict_like()
    169 elif is_list_like(arg):
    170     # we require a list, but not a 'str'
--> 171     return self.agg_list_like()
    173 if callable(arg):
    174     f = com.get_cython_func(arg)

File ~/pandas/pandas/core/apply.py:367, in Apply.agg_list_like(self)
    365 indices = []
    366 for index, col in enumerate(selected_obj):
--> 367     colg = obj._gotitem(col, ndim=1, subset=selected_obj.iloc[:, index])
    368     try:
    369         # Capture and suppress any warnings emitted by us in the call
    370         # to agg below, but pass through any warnings that were
    371         # generated otherwise.
    372         # This is necessary because of https://bugs.python.org/issue29672
    373         # See GH #43741 for more details
    374         with warnings.catch_warnings(record=True) as record:

File ~/pandas/pandas/core/window/rolling.py:289, in BaseWindow._gotitem(self, key, ndim, subset)
    284 if subset.ndim == 2 and (
    285     (is_scalar(key) and key in subset) or is_list_like(key)
    286 ):
    287     selection = key
--> 289 new_win = type(self)(subset, selection=selection, **kwargs)
    290 return new_win

File ~/pandas/pandas/core/window/rolling.py:141, in BaseWindow.__init__(self, obj, window, min_periods, center, win_type, axis, on, closed, method, selection)
    139 # TODO(2.0): Change this back to self.win_type once deprecation is enforced
    140 self._win_type = win_type
--> 141 self.axis = obj._get_axis_number(axis) if axis is not None else None
    142 self.method = method
    143 self._win_freq_i8 = None

File ~/pandas/pandas/core/generic.py:537, in NDFrame._get_axis_number(cls, axis)
    535     return cls._AXIS_TO_AXIS_NUMBER[axis]
    536 except KeyError:
--> 537     raise ValueError(f"No axis named {axis} for object type {cls.__name__}")

ValueError: No axis named 1 for object type Series

Expected Behavior

I expect the last rolling call to give a dataframe with the rolling sums and means of each row:

          0    1
0 sum   1.0  3.0
  mean  1.0  1.5
1 sum   3.0  7.0
  mean  3.0  3.5

Installed Versions

INSTALLED VERSIONS

commit : b4e578d
python : 3.9.10.final.0
python-bits : 64
OS : Darwin
OS-release : 21.3.0
Version : Darwin Kernel Version 21.3.0: Wed Jan 5 21:37:58 PST 2022; root:xnu-8019.80.24~20/RELEASE_X86_64
machine : x86_64
processor : i386
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.5.0.dev0+122.gb4e578db85
numpy : 1.22.1
pytz : 2021.3
dateutil : 2.8.2
pip : 22.0.3
setuptools : 60.8.2
Cython : 0.29.27
pytest : 7.0.0
hypothesis : None
sphinx : 4.4.0
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : 2.9.3
jinja2 : 3.0.3
IPython : 8.0.0
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
fsspec : 2022.01.0
gcsfs : None
matplotlib : 3.5.1
numba : None
numexpr : 2.8.1
odfpy : None
openpyxl : 3.0.9
pandas_gbq : 0.16.0
pyarrow : 6.0.1
pyreadstat : None
pyxlsb : None
s3fs : 2022.01.0
scipy : 1.7.3
sqlalchemy : 1.4.31
tables : 3.7.0
tabulate : None
xarray : 0.20.2
xlrd : 2.0.1
xlwt : None
zstandard : None

[xr-chen]

take

[sqali]

take

[sqali]

Is this still raising errors or fixed?

[anmyachev]

@jbrockmendel now I see another exception: NotImplementedError: axis other than 0 is not supported instead of ValueError.

I guess this is no longer a bug, but expected behavior. Should we close the issue?

[jbrockmendel]

I'd expect it to be deprecated, not raise.