CODEWITHSUNDEEP
pythonpandasdatetime
Ben Kailon
Ben Kailon

Reputation: 303

Convert/use timedelta's time to datetime

I have a dataframe with two columns: 1 timedelta 'Time', and 1 datetime 'DateTime'.

My timedelta column simply contains/displays a normal regular time, it never has more than 24 hours. It's not being used as a 'timedetla', just 'time'. It's just the way it comes when pandas gets the data from my database.

I want a new column 'NewDateTime', with the date from the datetime, and time from the deltatime.

So I have this:

        Time       DateTime     
1       09:01:00   2018-01-01 10:10:10
2       21:43:00   2018-01-01 11:11:11
3       03:20:00   2018-01-01 12:12:12

And I want this:

        Time       DateTime                NewDateTime
1       09:01:00   2018-01-01 10:10:10     2018-01-01 09:01:00
2       21:43:00   2018-01-01 11:11:11     2018-01-01 21:43:00
3       03:20:00   2018-01-01 12:12:12     2018-01-01 03:20:00

At first I tried to set the DateTime column's hours, minutes and seconds to 0. Then I planned to add the timedelta to the datetime.

But when I tried to do:

df['NewDateTime'] = df['DateTime'].dt.replace(hour=0, minute=0, second=0)

I get AttributeError: 'DatetimeProperties' object has no attribute 'replace'

Upvotes: 3

Views: 3770

Answers (1)

jezrael
jezrael

Reputation: 862441

Use Series.dt.floor for remove times:

df['NewDateTime'] = df['DateTime'].dt.floor('D') + pd.to_timedelta(df['Time'])
#if necesary convert times to strings
#df['NewDateTime'] = df['DateTime'].dt.floor('D') + pd.to_timedelta(df['Time'].astype(str))
print (df)
       Time            DateTime         NewDateTime
1  09:01:00 2018-01-01 10:10:10 2018-01-01 09:01:00
2  21:43:00 2018-01-01 11:11:11 2018-01-01 21:43:00
3  03:20:00 2018-01-01 12:12:12 2018-01-01 03:20:00

Upvotes: 4

Related Questions