Pandas dataframe combine two columns to get datetime column

Question

Given the dataframe below where uptime is time in nanoseconds, how to combine columns date_is with uptime to form a new column of type datetime object.

    date_is              uptime     
0   14/05/2016 10:54:33  11537640270059 
1   14/05/2016 10:54:33  11537650128140 
2   14/05/2016 10:54:33  11537659894659 
3   14/05/2016 10:54:33  11537679549779 
4   14/05/2016 10:54:33  11537699204899

jezrael · Accepted Answer · 2017-09-30 19:34:41Z

5

Use to_datetime + to_timedelta:

df['new'] = pd.to_datetime(df['date_is']) + pd.to_timedelta(df['uptime'])
print (df)
               date_is          uptime                           new
0  14/05/2016 10:54:33  11537640270059 2016-05-14 14:06:50.640270059
1  14/05/2016 10:54:33  11537650128140 2016-05-14 14:06:50.650128140
2  14/05/2016 10:54:33  11537659894659 2016-05-14 14:06:50.659894659
3  14/05/2016 10:54:33  11537679549779 2016-05-14 14:06:50.679549779
4  14/05/2016 10:54:33  11537699204899 2016-05-14 14:06:50.699204899

Is possible also converted columns assign back:

df['date_is'] = pd.to_datetime(df['date_is'])
df['uptime'] = pd.to_timedelta(df['uptime'])
df['new'] = df['date_is'] + df['uptime']
print (df)
              date_is          uptime                           new
0 2016-05-14 10:54:33 03:12:17.640270 2016-05-14 14:06:50.640270059
1 2016-05-14 10:54:33 03:12:17.650128 2016-05-14 14:06:50.650128140
2 2016-05-14 10:54:33 03:12:17.659894 2016-05-14 14:06:50.659894659
3 2016-05-14 10:54:33 03:12:17.679549 2016-05-14 14:06:50.679549779
4 2016-05-14 10:54:33 03:12:17.699204 2016-05-14 14:06:50.699204899

answered Sep 30, 2017 at 19:34

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

hebeha Over a year ago

This worked great! If your 'time' column comes in a different format, you first need to convert them to nanoseconds. Eg. my time column is hours in a military format like: 0 100 1 200 ... 22 2300 23 2400 and I simply multiplied it: pd.to_timedelta(df.Hour*100000*60*60) Ideally, we would pass the format as a parameter to to_timedelta ...

jezrael Over a year ago

@hebeha - You can use pd.to_timedelta(df.Hour, unit='h')

hebeha Over a year ago

with the military format I get each hour multiplied by 100 if I use unit='h', and to_timedelta doesn't have the string format, it would be '%-I' in this case.

hebeha Over a year ago

Another workaround would be 'pd.to_timedelta(data.Hour/100, unit='h')'

Collectives™ on Stack Overflow

Pandas dataframe combine two columns to get datetime column

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related