Pandas date_range inconsistent behavior

Question

When using the pandas (v. 13.1) date_range function I get inconsistent behavior about whether or not the 'end' is inside in the returned range:

In [1]: pd.date_range(start='2014-06-09 15:36:55', 
                      end='2014-06-09 15:37:46', 
                      freq='20s')
Out[1]: [2014-06-09 15:36:55, ..., 2014-06-09 15:37:55]

And

In [2]: pd.date_range(start='2014-06-09 15:36:55', 
                      end='2014-06-09 15:37:46', 
                      freq='10s')
Out[2]: [2014-06-09 15:36:55, ..., 2014-06-09 15:37:45]

Note that in the first case the last time is later that the specified 'end', while in the second case the last time is smaller. Can anyone explain this?

Answer 1

I believe this issue has already been fix in some release. Now the date_range behaviour is consistent, always including the start and the end of the DatetimeIndex always <= end. And the output format became pretty too.

Following code is using pandas 0.17.0:

In [1]: import pandas as pd

In [2]: pd.date_range(start='2014-06-09 15:36:55', 
   ...:                       end='2014-06-09 15:37:46', 
   ...:                       freq='20s')
Out[2]: 
DatetimeIndex(['2014-06-09 15:36:55', '2014-06-09 15:37:15',
               '2014-06-09 15:37:35'],
              dtype='datetime64[ns]', freq='20S')

In [3]: pd.date_range(start='2014-06-09 15:36:55', 
                      end='2014-06-09 15:37:46', 
                      freq='10s')
Out[3]: 
DatetimeIndex(['2014-06-09 15:36:55', '2014-06-09 15:37:05',
               '2014-06-09 15:37:15', '2014-06-09 15:37:25',
               '2014-06-09 15:37:35', '2014-06-09 15:37:45'],
              dtype='datetime64[ns]', freq='10S')

In [4]: pd.date_range(start='2014-06-09 15:36:55', 
                      end='2014-06-09 15:37:46', 
                      freq='15s')
Out[4]: 
DatetimeIndex(['2014-06-09 15:36:55', '2014-06-09 15:37:10',
               '2014-06-09 15:37:25', '2014-06-09 15:37:40'],
              dtype='datetime64[ns]', freq='15S')

In [5]: pd.date_range(start='2014-06-09 15:36:55', 
                      end='2014-06-09 15:37:46', 
                      freq='1s')
Out[5]: 
DatetimeIndex(['2014-06-09 15:36:55', '2014-06-09 15:36:56',
                   ......                    ......
               '2014-06-09 15:37:45', '2014-06-09 15:37:46'],
              dtype='datetime64[ns]', freq='S')

Note that the end of the DatetimeIndex never exceeds end date of date_range.