Feature Sugesstion: date_range from end to start

[npezolano]

When using pd.date_range we work our way forwards from start to end. It would be nice to pass a flag or use the dates in reverse to go from end to start when creating the date_range.

#EX:
pd.date_range(end, start)
#or
pd.date_range(start, end, end_to_start=True)

Ex:
start = datetime.datetime(2016 ,2, 8)
end = datetime.datetime(2018 , 6, 1)
pd.date_range(start end, freq='6m', start_to_end=True)
>>> DatetimeIndex(['2016-02-08', '2016-06-01', '2016-12-01', '2017-06-01',
               '2017-12-01', '2018-06-01'])

[max-sixty]

reversed?

In [7]: d=pd.date_range(start='2000', freq='Q', end='2002')

In [8]: d
Out[8]: 
DatetimeIndex(['2000-03-31', '2000-06-30', '2000-09-30', '2000-12-31',
               '2001-03-31', '2001-06-30', '2001-09-30', '2001-12-31'],
              dtype='datetime64[ns]', freq='Q-DEC')

In [9]: reversed(d)
Out[9]: <reversed at 0x1073aaa20>

In [10]: pd.Index(reversed(d))
Out[10]: 
DatetimeIndex(['2001-12-31', '2001-09-30', '2001-06-30', '2001-03-31',
               '2000-12-31', '2000-09-30', '2000-06-30', '2000-03-31'],
              dtype='datetime64[ns]', freq=None)

[npezolano]

Not reversed, the order doesn't change, but you work the algorithm from the end date to the start. I'll add an example above to make it more clear.

[jreback]

so pls provide a usecase. Generally you would simply reverse sort to work with things in the reverse order.

[rouxrei]

Here's a usecase:
You want to guarantee that your specified end_date is a value in the resulting set of values.

Say you do

start = datetime.datetime(2019 ,1, 1)
end = datetime.datetime(2019 , 6, 1)
pd.date_range(start, end, freq='4MS')

You get

DatetimeIndex(['2019-01-01', '2019-05-01'], dtype='datetime64[ns]', freq='4MS') 

Which means my specified end_date is not a value in my list.
(Note the month values)

With this flag, or "reverse generation" I would expect to get this instead:

start = datetime.datetime(2019 ,1, 1)
end = datetime.datetime(2019 , 6, 1)
pd.date_range(start, end, freq='4MS', reverse=True)

> DatetimeIndex(['2019-02-01', '2019-06-01'], dtype='datetime64[ns]', freq='4MS'))

(Note the month values)