apply seems to lose timezone awareness. Would expect the following to be invariant:
In [9]: import pandas as pd
In [10]: pd.__version__
Out[10]: u'0.17.1'
In [11]: ser = pd.Series([pd.Timestamp('2015/01/01', tz='UTC'), pd.Timestamp('2015/01/02', tz='UTC')])
In [12]: ser
Out[12]:
0 2015-01-01 00:00:00+00:00
1 2015-01-02 00:00:00+00:00
dtype: datetime64[ns, UTC]
In [13]: ser.apply(lambda x: x)
Out[13]:
0 2015-01-01
1 2015-01-02
dtype: datetime64[ns]
Comment From: jreback
xref #11564
cc @kawochen
yeh, .apply return inference is being revamped as of above.
Comment From: jreback
yep this is a dupe of #11757