[FIXED] Why is pd.date_range() sometimes stop-inclusive and sometimes exclusive based on the frequency?

Issue

import pandas as pd
print(pd.date_range(start='1999-08-01', end='2000-07-01', freq='D'))
print(pd.period_range(start='1999-08-01', end='2000-07-01', freq='D'))


print(pd.date_range(start='1999-08', end='2000-07', freq='M')) 
#This doesn't include anything from 2000-07, ie is not stop-inclusive - length 11


print(pd.period_range(start='1999-08', end='2000-07', freq='M')) # Length 12

Of these, three are stop-inclusive as one might naively define, whereas the date_range with frequency "M" is not. Why is that?

Solution

This is because 2000-07 is understood as 2000-07-01 and M is the end of month (opposite to MS the start of month) so 2000-07-31 is after 2000-07-01. If you want to include "July" change the frequency from M to MS but it will change the day:

>>> pd.date_range(start='1999-08', end='2000-07', freq='M')[[0, -1]]
DatetimeIndex(['1999-08-31', '2000-06-30'], dtype='datetime64[ns]', freq=None)

>>> pd.date_range(start='1999-08', end='2000-07', freq='MS')[[0, -1]]
DatetimeIndex(['1999-08-01', '2000-07-01'], dtype='datetime64[ns]', freq=None)

Answered By - Corralien