Problem with Series.clip with Timestamp threshold

[miketkelly]

>>> import pandas as pd
>>> s = pd.Series(pd.date_range('2013-05-01', '2013-05-03'))
>>> s
0   2013-05-01 00:00:00
1   2013-05-02 00:00:00
2   2013-05-03 00:00:00
dtype: datetime64[ns]

>>> s.clip_lower(s[1])
0    2013-05-02 00:00:00
1    1367452800000000000
2    1367539200000000000
dtype: object

>>> s.clip_lower(s[1].value)
0   2013-05-02 00:00:00
1   2013-05-02 00:00:00
2   2013-05-03 00:00:00
dtype: datetime64[ns]

>>> print pd.__version__
0.11.1.dev-f1a3226

Any reason not to implement clip_lower and clip_upper with Series.where rather than np.where? It solves this problem.

[jreback]

yes....Serise.where knows how to deal with datetime64[ns] and NaT so that would be a good change

if you want to do a PR, pls add a test for this (and see that existing tests pass)

thanks!