How to get timestamp differences per group?

April 10, 2024

I have a column date in my dataset and column group.i want to get parameter differrence between min and max date in "date" column per group. how to do that? here i example of my data:

group      date
main      2024-01-01
main      2024-01-03
main      2024-01-05
second    2024-02-05
second    2024-02-20

desire result:

group      date_diff
main          4
second        15

i tried to do groupby and then i dont know how to get differences

>Solution :

Get minimal and maximal datetimes per groups by GroupBy.agg, then get difference by subtract and converting days to numbers by Series.dt.days:

df['date'] = pd.to_datetime(df['date'])

out = (df.groupby('group')['date'].agg(['min','max'])
         .assign(date_diff=lambda x: x['max'].sub(x['min']).dt.days)['date_diff']
         .reset_index())

print (out)
    group  date_diff
0    main          4
1  second         15

Or use lambda function, if large DataFrame, should be slowier:

out = (df.groupby('group')['date'].agg(lambda x: (x.max() - x.min()).days)
         .reset_index(name='date_diff'))
print (out)
    group  date_diff
0    main          4
1  second         15