Inconsistent behaviour with min method on object dtype columns

[turner3467]

xref #18021, #16832

Code Sample, a copy-pastable example if possible

df = pd.DataFrame({'col1': [0, 1, 2, 3], 'col2': ['a', 'b', None, 'd'], 'col3': ['e', 'f', None, 'h']})
df.min(axis=0, skipna=True)

Problem description

If I've read the documentation correctly I think min should return a series of the same length as the relevant axis of the dataframe, and especially with the skipna flag set as True (which is the default) NA values should be ignored in the calculation.

Output

col1 0
dtype: int64

Expected Output

col1 0
col2 a
col3 e
dtype: object

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.1.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 78 Stepping 3, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.20.3
pytest: 3.0.7
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.25.2
numpy: 1.12.1
scipy: 0.19.0
xarray: None
IPython: 5.3.0
sphinx: 1.5.6
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2017.2
blosc: None
bottleneck: 1.2.1
tables: 3.2.2
numexpr: 2.6.2
feather: None
matplotlib: 2.0.2
openpyxl: 2.4.7
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.3
bs4: 4.6.0
html5lib: 0.999
sqlalchemy: 1.1.9
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
pandas_gbq: None
pandas_datareader: None

[jreback]

so couple of things going on.

.min(numeric_only=None) by default so non-numerics are ignored (if they raise). But these fail anyhow if executed directly on the column.

In [8]: df.col2.min()
TypeError: '<=' not supported between instances of 'str' and 'float'

so we are not skipping NaN here and passing directly to numpy which raises (would have to patch nanmax as well);

In [1]: df = pd.DataFrame({'col1': [0, 1, 2, 3], 'col2': ['a', 'b',None, 'd'], 'col3': ['e', 'f', None, 'h']})

In [2]: df.col2.min()
Out[2]: 'a'

In [3]: df.min()
Out[3]: 
col1    0
col2    a
col3    e
dtype: object

patch

(pandas) bash-3.2$ git diff 
diff --git a/pandas/core/nanops.py b/pandas/core/nanops.py
index e1c09947a..33377002e 100644
--- a/pandas/core/nanops.py
+++ b/pandas/core/nanops.py
@@ -478,6 +478,8 @@ def _nanminmax(meth, fill_value_typ):
             except:
                 result = np.nan
         else:
+
+            values = values[~mask]
             result = getattr(values, meth)(axis)
 
         result = _wrap_results(result, dtype)

[jreback]

so open to fixing this (PR welcome!) it might break some existing tests.

[TomAugspurger]

#31757 was closed, so moving this off 1.0.2