Closed
Description
Several of the functions in nanops
, including nanmean
, nanmedian
, and anything using _get_counts
will crash when using a 1-dimensional all-nan array when the axis
argument is set to 0.
>>> from pandas.core import nanops
>>> import bottleneck
>>> import numpy as np
>>> nanops._USE_BOTTLENECK = False
>>>
>>> val = np.tile(np.nan, (5,))
>>>
>>> bottleneck.nanmean(val, axis=0)
nan
>>> nanops.nanmean(val, axis=0)
TypeError: 'numpy.float64' object does not support item assignment