Converting Zulu Date/Time to Just Date Using Python and Pandas-CodePudding

I have start and end times that look as follows:

Start: 2017-03-05T19:18:53Z

End: 2017-03-05T19:57:54.042000Z

Current iteration that leaves me with the right format but the wrong dtype:

format_time_user_df[['start', 'end']] = format_time_user_df[['start', 'end']].apply(pd.to_datetime, errors='coerce')
cols = ['start', 'end']
format_time_user_df[cols] = format_time_user_df[cols].apply(lambda x: pd.to_datetime(x, unit='ms').dt.date)

resulting in the start and end times that look as follows:

Start: 2017-03-05

End: 2017-03-05

but these are now objects and not dates.

What I am trying to accomplish is to get the dates into ascending order and also, if there is a way to section of a week at a time (Monday through Sunday) for data analysis.

CodePudding user response：

If you use .loc[:, ...] =, your values will be updated but the column not, so:

Replace:

df.loc[:, ['start', 'end']] = ...

By:

df[['start', 'end']] = ...

Now you can use:

>>> df['start'].dt.normalize()
0   2017-03-05 00:00:00 00:00
Name: start, dtype: datetime64[ns, UTC]

Setup:

df = pd.DataFrame({'start': ['2017-03-05T19:18:53Z'],
                   'end': ['2017-03-05T19:57:54.042000Z']})