Python Pandas df, subtract minutes (in time duration column) from time column

I have a pandas df with Time, Duration (in minutes, integer format), and Duration in timedelta format. The EndTime column has been parsed to time format.

I wanted to created a new column: StartTime

I tried various things: from datetime import timedelta

 df['StartTimeEstimate'] = df['EndTime'] - pd.Timedelta(hours=0, minutes=df['Minutes'], seconds=0)

or

df['StartTimeEstimate'] = df['EndTime'] - timedelta(minutes = df['Minutes'])

But couldn't quite work out a StartTime.

Ideally, I would like to see end StartTime to be: 22:53:18 12:29:38 04:59:38

You can't add time and timedelta, as the former doesn't have information about the date.

So if 'EndTime' is of type time, you can first convert it to str and use pd.to_datetime to convert it to datetime. Then you can just subtract the timedelta from the datetime:

df["StartTimeEstimate"] = (
    pd.to_datetime(df["EndTime"].astype(str)) - df["Duration_TimeDelta"]
).dt.time

   id   EndTime  Duration_Minutes Duration_TimeDelta StartTimeEstimate
0   1  00:53:18               120    0 days 02:00:00          22:53:18
1   2  01:04:38                35    0 days 00:35:00          00:29:38
2   3  05:04:38                 5    0 days 00:05:00          04:59:38

If 'EndTime' is of type str, you can remove the .astype(str).