series.agg(np.nansum) etc. returns a series, scalar expected

[topper-123]

>>> pd.Series([1,2,3,4]).agg(np.sum)
10
>>> pd.Series([1,2,3,4]).agg(np.nansum)
0    1
1    2
2    3
3    4
dtype: int64

I would expect the return values of .agg(np.nansum) and .agg(np.sum) to be the same, given that pandas "normalizes" np.sum into np.nansum (don't know the correct term).

the same issue is with np.nanmean, np.nanmax etc.

Problem description

The underlying issue is that SelectionMixIn._cython_table has records for np.sum, np.mean etc, but not np.nansum, np.nanmean etc.

From a quick look at it, it seems that the issue is easily fixable, by adding the np.nan* function to cython_table.

Output of pd.show_versions()

INSTALLED VERSIONS

commit: 7590b87
python: 3.6.3.final.0
python-bits: 32
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: 0.22.0.dev0+710.g7590b87.dirty
pytest: 3.3.1
pip: 9.0.1
setuptools: 38.2.5
Cython: 0.26.1
numpy: 1.13.3
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.2.1
sphinx: 1.6.3
patsy: 0.4.1
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.1.0
openpyxl: 2.4.9
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: None
lxml: None
bs4: None
html5lib: 1.0b10
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[jreback]

we don't have the nan functions in our _cython_map It think. welcome to a PR

I see you said that!

[AaronCritchley]

Made an attempt in this PR but couldn't quite get it to work, hopefully somebody else is able to pick this up. I'll try again at a later date if not 😄