Convert object year-week to datetime format-CodePudding

I have the data as below:

    New_Time    Visitors
0   2016-01 4.709530
1   2016-02 4.219508
2   2016-03 4.094345
3   2016-04 4.488636
4   2016-05 4.584967

New_time format is Year-WeekNumber, but it's currently held as dtype 'object'. I'm trying to convert New_Time from object to datetime dtype, e.g.

df_weekly_vistors_logscale['New_Time'] = df_weekly_vistors_logscale['New_Time'].dt.week

df_weekly_vistors_logscale['New_Time'] = pd.to_datetime(df_weekly_vistors_logscale['New_Time'], format = '%Y%V')

df_weekly_vistors_logscale['New_Time'] = df_weekly_vistors_logscale['New_Time'].dt.strftime('%Y-%W')

None of which are working with various errors. I'm sure the problem is logical but can't figure this one out.

Any help?

Thank you.

CodePudding user response：

Try below

import datetime
def to_datetime(x):
  return datetime.datetime.strptime(x   '-1', "%Y-%W-%w")

df_weekly_vistors_logscale['New_Time'] = df_weekly_vistors_logscale['New_Time'].apply(to_datetime)

CodePudding user response：

Try this:

import pandas as pd
df_weekly_vistors_logscale['New_Time'] = pd.to_datetime(df_weekly_vistors_logscale['New_Time'].dt.strftime('%Y-%W'))

CodePudding user response：

as an addition to the linked duplicate, if you want to convert ISO week, you can alternatively use

>>> import pandas as pd
>>> s = pd.Series(["2016-01", "2016-02", "2016-03"])
>>> pd.to_datetime(s.add("-1"), format="%G-%V-%u")
0   2016-01-04
1   2016-01-11
2   2016-01-18
dtype: datetime64[ns]
>>>