Summing with missing data appears to be broken in 0.16.2 when all elements are NA

[numpand]

Summing with missing data appear to be broken in 0.16.2. According to http://pandas.pydata.org/pandas-docs/version/0.16.2/missing_data.html#calculations-with-missing-data
When summing data, NA (missing) values will be treated as zero
=>>> If the data are all NA, the result will be NA <== This returns zero in 0.16.2 (it returns NA in 0.15.2).
Old behavior as specified in documentation:

New behavior which is different from documenation:

[jreback]

You must have a confused environment. Both 0.15.2 & 0.16.2 have the same results. Note you may be using a different numpy.

FYI, pls don't paste images, they are impossible to copy-paste.

In [1]: df = DataFrame([[np.nan,np.nan],[np.nan,1],[2,3]],columns=list('AB'))

In [2]: df
Out[2]: 
    A   B
0 NaN NaN
1 NaN   1
2   2   3

In [3]: df.sum(axis=1)
Out[3]: 
0   NaN
1     1
2     5
dtype: float64

In [4]: np.__version__
Out[4]: '1.9.2'

In [5]: pd.__version__
Out[5]: '0.15.2-319-g434828a'

In [1]: df = DataFrame([[np.nan,np.nan],[np.nan,1],[2,3]],columns=list('AB'))

In [2]: df.sum(axis=1)
Out[2]: 
0   NaN
1     1
2     5
dtype: float64

In [3]: pd.__version__
Out[3]: '0.16.2'

[numpand]

Sorry about the images. I just checked the versions and we are using numpy 1.9.2 with pandas 0.16.2 and yet still getting a zero when summing all NAs. Any idea why?

Python 3.3.5 |Continuum Analytics, Inc.| (default, Jan 9 2015, 10:46:22) [MSC v.1600 32 bit (Intel)]
Type "copyright", "credits" or "license" for more information.

IPython 3.2.1 -- An enhanced Interactive Python.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
? -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help -> Python's own help system.
object? -> Details about 'object', use 'object??' for extra details.

In [1]: import pandas as pd

In [2]: import numpy as np

In [3]: pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.3.5.final.0
python-bits: 32
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 60 Stepping 3, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.16.2
nose: 1.3.4
Cython: 0.22.1
numpy: 1.9.2
scipy: 0.15.1
statsmodels: 0.6.1
IPython: 3.2.1
sphinx: 1.3.1
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.4
bottleneck: 1.0.0
tables: 3.1.1
numexpr: 2.3.1
matplotlib: 1.4.3
openpyxl: 1.8.5
xlrd: 0.9.3
xlwt: 1.0.0
xlsxwriter: 0.7.3
lxml: 3.4.4
bs4: 4.3.2
html5lib: 0.999
httplib2: None
apiclient: None
sqlalchemy: 1.0.6
pymysql: None
psycopg2: 2.5.5 (dt dec pq3 ext)

In [4]: df = pd.DataFrame([[np.nan,np.nan],[np.nan,1],[2,3]],columns=list('AB'))

In [5]: df.sum(axis=1)
Out[5]:
0 0
1 1
2 5
dtype: float64

[shoyer]

This is a dup of #9422