Python creating a new dataframe with list of unique dates as index

Question

I have dateframe with index as df.index =

2016-08-01 06:45:00    
2016-08-01 07:00:00    
2016-08-01 07:15:00    
.
.
2018-03-28 11:30:00    
2018-03-28 11:45:00    
2018-03-28 12:00:00  

I want to create a new dataframe that it will have only unique dates as

new_df.index =

2016-08-01
2016-08-02
.
.
2018-03-28
2018-03-29

So, how to create a new dataframe with unique dates as index?

Answer 1

There is a build-in collection set in Python, that have only unique elements. You can do something like it:

new_data = sorted(list(set(old_data)))

If you want to crop out time in your datetime lines, you can modify this code with generators:

new_data = sorted(list(set([elem[:10] for elem in old_data])))

Note, that if you have some info linked with your datetimes (like values in dict with datetime keys), you must handle elements removal before doing it.

Answer 2

Since you have not provided any data I will assume it is not important.

Your first DataFrame appears to have a DatetimeIndex and it appears you want to convert it to a PeriodIndex . You can do this to get the unique days with df.resample(rule='D').asfreq() . Docs are here