Series.str.contains doesn't correct if series contains only nan values.

[ponomarevvl90]

Code Sample, a copy-pastable example if possible

import pandas as pd
import numpy as np
import re

CONTAINS_REG_EXP = ur'{0}'
value = '3'
t1 = pd.Series(data=[np.nan, np.nan, np.nan], dtype=np.object)
t2 = t1.str.contains(CONTAINS_REG_EXP.format(re.escape(value)),
                                  flags=re.U, na=False)
print t2

Expected Output

0    False
1    False
2    False
dtype: bool

output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 2.7.10.final.0
python-bits: 64
OS: Linux
OS-release: 4.2.0-42-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.17.1
nose: None
pip: 8.0.3
setuptools: 18.0.1
Cython: 0.23.4
numpy: 1.10.1
scipy: None
statsmodels: None
IPython: 4.0.1
sphinx: None
patsy: None
dateutil: 2.4.2
pytz: 2015.7
blosc: None
bottleneck: None
tables: 3.2.2
numexpr: 2.4.6
matplotlib: None
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: 2.6.1 (dt dec pq3 ext lo64)
Jinja2: None

[jreback]

Here's a simpler example. This is casting only when na flag is passed.

In [20]: Series([np.nan, np.nan, np.nan],dtype='object').str.contains('foo')
Out[20]: 
0   NaN
1   NaN
2   NaN
dtype: float64

In [21]: Series([np.nan, np.nan, np.nan],dtype='object').str.contains('foo',na=False)
Out[21]: 
0    0.0
1    0.0
2    0.0
dtype: float64

In [22]: Series([np.nan, np.nan, np.nan],dtype='object').str.contains('foo',na=True)
Out[22]: 
0    1.0
1    1.0
2    1.0
dtype: float64

works fine if not all-nan object dtype.

In [23]: Series([np.nan, np.nan, np.nan, 'a'],dtype='object').str.contains('foo')
Out[23]: 
0      NaN
1      NaN
2      NaN
3    False
dtype: object

In [25]: Series([np.nan, np.nan, np.nan, 'a'],dtype='object').str.contains('foo', na=False)
Out[25]: 
0    False
1    False
2    False
3    False
dtype: bool

In [26]: Series([np.nan, np.nan, np.nan, 'a'],dtype='object').str.contains('foo', na=True)
Out[26]: 
0     True
1     True
2     True
3    False
dtype: bool

[jreback]

pull-requests to fix are welcome!

[josh-howes]

First time contributor here. I'll take this on.