Pandas: Take date from one column and add it to other

by

There is dataframe as:

time         date    
11:42:14.556 2023-12-10 10:12:00.129465868
11:42:16.005 2023-12-11 15:25:00.759067848
11:42:32.175 2023-12-12 13:56:00.653087940

Columns are object types.

How to take date from “date” column and add it to “time” column?

  • Can you clarify what are the dtype of the original columns?

    – 

  • The type of data determines the method.

    – 

  • All columns are object types

    – 

  • Thanks @harp1814 for clarifying, then both my approaches should work 🙂

    – 

  • @PandaKim the nice thing with to_datetime is that it can take a datetime column as input, so in doubt the conversion doesn’t hurt.

    – 

Use dt.normalize after converting to_datetime:

df['time'] = pd.to_datetime(df['date']).dt.normalize().add(pd.to_timedelta(df['time']))

Or if both columns are strings, extract the date and concatenate:

df['time'] = df['date'].str.extract('(\S+)', expand=False)+' '+df['time']

# or
df['time'] = df['date'].str[:10]+' '+df['time']

Output:

                     time                           date
0 2023-12-10 11:42:14.556  2023-12-10 10:12:00.129465868
1 2023-12-11 11:42:16.005  2023-12-11 15:25:00.759067848
2 2023-12-12 11:42:32.175  2023-12-12 13:56:00.653087940

Use Series.str.split with add times and convert output to datetimes:

df['time'] = pd.to_datetime(df['date'].str.split().str[0] + ' ' + df['time'].astype(str))

Or Series.dt.normalize with to_timedelta:

df['time'] = df['date'].dt.normalize() + pd.to_timedelta(df['time'])

print (df)
                     time                           date
0 2023-12-10 11:42:14.556  2023-12-10 10:12:00.129465868
1 2023-12-11 11:42:16.005  2023-12-11 15:25:00.759067848
2 2023-12-12 11:42:32.175  2023-12-12 13:56:00.653087940

Leave a Comment