set_index does not set 'freq'

[joshplumley]

pandas 0.17.0 does not seem to set the freq attribute on the index anymore. I believe 0.16.2 did.

Example:

import pandas
import numpy
import datetime

df = pandas.DataFrame(numpy.random.randn(24, 1), columns=['a'], index=pandas.date_range(datetime.datetime(2015, 10, 1), datetime.datetime(2015,10,1,23), freq='H', tz='US/Eastern'))
new_index = pandas.date_range(datetime.datetime(2015, 10, 2), datetime.datetime(2015,10,2,23), freq='H', tz='US/Eastern')
new_df = df.set_index(new_index)

In [2]: df.index
Out[2]: 
DatetimeIndex(['2015-10-01 00:00:00-04:00', '2015-10-01 01:00:00-04:00',
               '2015-10-01 02:00:00-04:00', '2015-10-01 03:00:00-04:00',
               '2015-10-01 04:00:00-04:00', '2015-10-01 05:00:00-04:00',
               '2015-10-01 06:00:00-04:00', '2015-10-01 07:00:00-04:00',
               '2015-10-01 08:00:00-04:00', '2015-10-01 09:00:00-04:00',
               '2015-10-01 10:00:00-04:00', '2015-10-01 11:00:00-04:00',
               '2015-10-01 12:00:00-04:00', '2015-10-01 13:00:00-04:00',
               '2015-10-01 14:00:00-04:00', '2015-10-01 15:00:00-04:00',
               '2015-10-01 16:00:00-04:00', '2015-10-01 17:00:00-04:00',
               '2015-10-01 18:00:00-04:00', '2015-10-01 19:00:00-04:00',
               '2015-10-01 20:00:00-04:00', '2015-10-01 21:00:00-04:00',
               '2015-10-01 22:00:00-04:00', '2015-10-01 23:00:00-04:00'],
              dtype='datetime64[ns, US/Eastern]', freq='H')

In [3]: new_df.index
Out[3]: 
DatetimeIndex(['2015-10-02 00:00:00-04:00', '2015-10-02 01:00:00-04:00',
               '2015-10-02 02:00:00-04:00', '2015-10-02 03:00:00-04:00',
               '2015-10-02 04:00:00-04:00', '2015-10-02 05:00:00-04:00',
               '2015-10-02 06:00:00-04:00', '2015-10-02 07:00:00-04:00',
               '2015-10-02 08:00:00-04:00', '2015-10-02 09:00:00-04:00',
               '2015-10-02 10:00:00-04:00', '2015-10-02 11:00:00-04:00',
               '2015-10-02 12:00:00-04:00', '2015-10-02 13:00:00-04:00',
               '2015-10-02 14:00:00-04:00', '2015-10-02 15:00:00-04:00',
               '2015-10-02 16:00:00-04:00', '2015-10-02 17:00:00-04:00',
               '2015-10-02 18:00:00-04:00', '2015-10-02 19:00:00-04:00',
               '2015-10-02 20:00:00-04:00', '2015-10-02 21:00:00-04:00',
               '2015-10-02 22:00:00-04:00', '2015-10-02 23:00:00-04:00'],
              dtype='datetime64[ns, US/Eastern]', freq=None)

The datetime appears to have been set correctly, but the freq is = None on the new_df. Is this normal behavior with pandas 0.17.0?

[jreback]

this was fixed here: #11301, and will be in 0.17.1 (was an unrelated bug-report)