Confusing - rolling min( ) function "accepting" objects

[gojandrooo]

Code Sample, a copy-pastable example if possible

s = pd.Series([3,1,2,'x','y','z'])

s.rolling(2).min()

0    3
1    1
2    2
3    x
4    y
5    z
dtype: object

Problem description

Should raise an exception because the series contains objects. Instead, function returns the series.

[jorisvandenbossche]

Thanks for the report!

This is certainly a bug, as it is not even calculating a minimum for the numerical values. Didn't look in detail, but I suppose this is due to the automatic skipping of non-numerical columns.
In case of a DataFrame that is useful:

In [3]: pd.DataFrame({'a': [1,2,3,4], 'b':[1,2,'a', 'b']}).rolling(2).min()
Out[3]: 
     a  b
0  NaN  1
1  1.0  2
2  2.0  a
3  3.0  b

but in case of a Series, just passing it through makes less sense.

[jorisvandenbossche]

Related issues: #18129, #15085

[ihsansecer]

This is working fine on master

>>> pd.DataFrame({'a': [1,2,3,4], 'b':[1,2,'a', 'b']}).rolling(2).min()
     a
0  NaN
1  1.0
2  2.0
3  3.0

>>> s = pd.Series([3,1,2,'x','y','z'])
>>> s.rolling(2).min()
Traceback (most recent call last):
...
pandas.core.base.DataError: No numeric types to aggregate

[jreback]

#27044 was just merged