Issue passing a Timestamp to Series.between_time()

[SimonFransen]

Hi,

I have an issue with passing pandas.Timestamp objects as arguments to pandas.Series.between_time(start, end).

It appears that pandas.tseries.index.DatetimeIndex.indexer_between_time() tries to convert start_time and end_time arguments as strings.

The following lines give a temporary fix:

if not isinstance(start_time, Timestamp):
        start_time = to_time(start_time)
if not isinstance(end_time, Timestamp):
        end_time = to_time(end_time)

Thanks,
Simon

[jreback]

pls pd.show_versions()
pls show what the actual issue is via code sample

this was changed very explicitly for 0.18.0

[jreback]

dupe of #11818, closed by #11832

In [12]: s = Series(np.arange(10),index=pd.date_range('20130101 09:00',periods=10,freq='H'))

In [13]: s
Out[13]: 
2013-01-01 09:00:00    0
2013-01-01 10:00:00    1
2013-01-01 11:00:00    2
2013-01-01 12:00:00    3
2013-01-01 13:00:00    4
2013-01-01 14:00:00    5
2013-01-01 15:00:00    6
2013-01-01 16:00:00    7
2013-01-01 17:00:00    8
2013-01-01 18:00:00    9
Freq: H, dtype: int64

In [14]: s.between_time(Timestamp('20130101 10:00'),Timestamp('20130101 14:00'))
ValueError: Cannot convert arg [Timestamp('2013-01-01 10:00:00')] to a time

In [15]: s.between_time('10:00','14:00')
Out[15]: 
2013-01-01 10:00:00    1
2013-01-01 11:00:00    2
2013-01-01 12:00:00    3
2013-01-01 13:00:00    4
2013-01-01 14:00:00    5
Freq: H, dtype: int64