[Python] Select rows from dataframe by group by and select max in given column

I have a data set structured as follow. This a sample so please imagine a dataframe which contain many sequences. Few things to precise: time are in descending order. created_at is also in descending order. Both time and created_at reset when a new source OR currency comes up.

index time app_v last source currency created_at
1 2019-10-23 3 4488 gol JPY 66
2 2019-10-23 29193 6687 gol JPY 65
3 2019-10-22 3 4587 gol JPY 64
4 2019-10-20 3 5687 gol JPY 63
5 2019-10-19 3 34787 gol JPY 62
6 2019-10-18 3 11789 gol JPY 61
...


The expected output is the latest last in terms of time grouped by source currency and app_v. For this sample, we have 2 rows corresponding:

index time app_v last source currency created_at
1 2019-10-23 3 4488 gol JPY 66
2 2019-10-23 29193 6687 gol JPY 65

Continue reading...