How can I construct an elapsed time column from a column of pandas.datetime64?

Question

I have a dataframe containing a column with:

3    3/20/2017 8:00:36 PM
4    3/20/2017 8:01:09 PM
5    3/20/2017 8:01:43 PM
Name: ActualTime, dtype: object

I want to create a column with the elapsed time in seconds.

0
33
67

Tried things like:

t0 = pd.to_datetime(x,unit='s')
t1 = pd.to_datetime(y,unit='s')
t0-t1

3   NaT
4   NaT
Name: ActualTime, dtype: timedelta64[ns]

score 1 · Accepted Answer · answered Apr 17 '17 at 21:12

Here is what I ended up using:

df['ElapsedTime'] = (df['ActualTime'] - df['ActualTime'].iloc[0]).astype("timedelta64[m]")

I ended up using minutes instead for convenience. For seconds:

df['ElapsedTime'] = (df['ActualTime'] - df['ActualTime'].iloc[0]).astype("timedelta64[s]")

1 Answers1