Error when using total_seconds to subtract two dates with different formats in Python

I have two different dates and I want to know the difference in seconds between them.

My data:

id              date1                           date2                    date3
 1      2017-03-03T02:38:30.544000        2017-03-03T02:30:21      2017-03-03T02:45:21

I want to add two column to my data frame:

date_first_sec= date1-date2
date_last_sec=  date3-date2

My code:

import datetime as dt

df.date1= dt.datetime(df.date1)
df.date2= dt.datetime(df.date2)
df.date3= dt.datetime(df.date3)

df['date_first_sec']= (df.date1- df.date2).total_seconds()
df['date_last_sec']= (df.date3- df.date2).total_seconds() 

This is the error:

TypeError: cannot convert the series to <class 'int'>

Answer

Use .dt.total_seconds()

Ex:

df = pd.DataFrame([{"date1": "2017-03-03T02:38:30.544000", "date2":"2017-03-03T02:30:21", 'date3': "2017-03-03T02:45:21"}])

df.date1= pd.to_datetime(df.date1)
df.date2= pd.to_datetime(df.date2)
df.date3= pd.to_datetime(df.date3)

df['date_first_sec']= (df.date1 - df.date2).dt.total_seconds()
df['date_last_sec']= (df.date3 - df.date2).dt.total_seconds()
print(df)

Output:

                    date1               date2  ... date_first_sec  date_last_sec
0 2017-03-03 02:38:30.544 2017-03-03 02:30:21  ...        489.544          900.0

[1 rows x 5 columns]