REGR: unhelpful error message with np.min on unordered Categorical

[simonjayhawkins]

xref #28949

on master

>>> import numpy as np
>>> import pandas as pd
>>>
>>> pd.__version__
'1.1.0.dev0+1008.g60b0e9fbc'
>>>
>>> cat = pd.Categorical(["a", "b", "c", "b"], ordered=False)
>>> cat
[a, b, c, b]
Categories (3, object): [a, b, c]
>>>
>>> np.min(cat)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<__array_function__ internals>", line 6, in amin
  File "C:\Users\simon\Anaconda3\envs\pandas-dev\lib\site-packages\numpy\core\fromnumeric.py", line 2793, in amin
    keepdims=keepdims, initial=initial, where=where)
  File "C:\Users\simon\Anaconda3\envs\pandas-dev\lib\site-packages\numpy\core\fromnumeric.py", line 88, in _wrapreduction
    return reduction(axis=axis, out=out, **passkwargs)
  File "C:\Users\simon\pandas\pandas\util\_decorators.py", line 212, in wrapper
    return func(*args, **kwargs)
TypeError: min() got an unexpected keyword argument 'axis'

on 0.25.3

>>> import numpy as np
>>> import pandas as pd
>>>
>>> pd.__version__
'0.25.3'
>>>
>>> cat = pd.Categorical(["a", "b", "c", "b"], ordered=False)
>>> cat
[a, b, c, b]
Categories (3, object): [a, b, c]
>>>
>>> np.min(cat)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\simon\Anaconda3\lib\site-packages\numpy\core\fromnumeric.py", line 2618, in amin
    initial=initial)
  File "C:\Users\simon\Anaconda3\lib\site-packages\numpy\core\fromnumeric.py", line 84, in _wrapreduction
    return reduction(axis=axis, out=out, **passkwargs)
  File "C:\Users\simon\Anaconda3\lib\site-packages\pandas\core\arrays\categorical.py", line 2278, in min
    self.check_for_ordered("min")
  File "C:\Users\simon\Anaconda3\lib\site-packages\pandas\core\arrays\categorical.py", line 1586, in check_for_ordered
    "Categorical to an ordered one\n".format(op=op)
TypeError: Categorical is not ordered for operation min
you can use .as_ordered() to change the Categorical to an ordered one

[simonjayhawkins]

regression caused by #27929 (between 0.25.3 and 1.0)

37526c1 is the first bad commit
commit 37526c1
Author: Mak Sze Chun makszechun@gmail.com
Date: Mon Dec 2 18:37:11 2019 +0800

API/DEPR: Change default skipna behaviour + deprecate numeric_only in Categorical.min and max (#27929)

cc @makbigc

[topper-123]

I'm leaning on that this is more a bug in numpy.

numpy does in core/fromnumeric.py::_wrapreduction:

    return reduction(axis=axis, out=out, **passkwargs)

i.e. forcing parameters axis=None and out=None down to our function, requiring us to have those parameters also. Seems to make better sense to push axis and out into the passkwargs dict, but drop them there if they're None so downstream would not require axis and out parameters?

Of course that's beyond pandas to fix (and I may misunderstand _wrapreduction of course), and I agree that np.min(cat) should work, so I'm ok with doing a fix for this in pandas.