Is there a way to calculate the running total across only a few columns (unique values only)?

Question

I am trying to calculate the running total across a few specific columns of my dataFrame and I am only interested in calculating using unique values.

I have below an example dataframe:

Name	Product	Date	Location	Type	Sales	Ship Fee %	Total Fee
Tom	Bananas	01-01-2021	NY	Fruit	120	0.01	1.2
Tom	Apples	01-01-2021	NY	Fruit	120	0.01	1.2
Tom	Bananas	02-01-2021	TX	Fruit	420	0.01	4.2
Tom	Bananas	02-01-2021	TX	Fruit	120	0.01	1.2
Mat	Bananas	02-01-2021	NY	Fruit	30	0.01	0.3

I want to have a Running Total column, but only considering the Name and Date (as groupBy columns) and showing the sum of the unique values of Total Fee column. That would result in something like this:

Name	Product	Date	Location	Type	Sales	Ship Fee %	Total Fee	Running Total
Tom	Bananas	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
Tom	Apples	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
Tom	Bananas	02-01-2021	TX	Fruit	420	0.01	4.2	4.2
Tom	Bananas	02-01-2021	TX	Fruit	120	0.01	1.2	5.4
Mat	Bananas	02-01-2021	NY	Fruit	30	0.01	0.3	0.3

I am lost -> I haven’t been able to find anything that can give me this result.

Asked By: pipocaDourada

||

Source

Answer 1

I think this is what you are looking for:

Option 1: groupby "Name" and "Date" then cumsum only unique values for Total Fee

df['Running Total'] = df.drop_duplicates(['Name', 'Date', 'Total Fee']).groupby(['Name', 'Date'])['Total Fee'].cumsum()
df['Running Total'] = df['Running Total'].fillna(df['Total Fee'])

Option 2: groupby "Name", "Product", "Date". Then cumsum –> gives the accumulated sum for each product on each day for each person.

df['Running Total'] = df.groupby(['Name', 'Product','Date'], as_index=False)['Total Fee'].cumsum()

Testing and examples

Given this dataframe:

	Name	Product	Date	Location	Type	Sales	Ship Fee %	Total Fee
0	Tom	Bananas	01-01-2021	NY	Fruit	120	0.01	1.2
1	Tom	Apples	01-01-2021	NY	Fruit	120	0.01	1.2
2	Tom	Bananas	02-01-2021	TX	Fruit	420	0.01	4.2
3	Tom	Bananas	02-01-2021	TX	Fruit	120	0.01	1.2
4	Mat	Bananas	02-01-2021	NY	Fruit	30	0.01	0.3
5	Mat	Bananas	02-01-2021	NY	Fruit	50	0.01	0.3
6	Mat	Apples	03-01-2021	NY	Vegetable	80	0.02	1.6

Option 1 result:

	Name	Product	Date	Location	Type	Sales	Ship Fee %	Total Fee	Running Total
0	Tom	Bananas	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
1	Tom	Apples	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
2	Tom	Bananas	02-01-2021	TX	Fruit	420	0.01	4.2	4.2
3	Tom	Bananas	02-01-2021	TX	Fruit	120	0.01	1.2	5.4
4	Mat	Bananas	02-01-2021	NY	Fruit	30	0.01	0.3	0.3
5	Mat	Bananas	02-01-2021	NY	Fruit	50	0.01	0.3	0.3
6	Mat	Apples	03-01-2021	NY	Vegetable	80	0.02	1.6	1.6

Option 2 result:

	Name	Product	Date	Location	Type	Sales	Ship Fee %	Total Fee	Running Total
0	Tom	Bananas	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
1	Tom	Apples	01-01-2021	NY	Fruit	120	0.01	1.2	1.2
2	Tom	Bananas	02-01-2021	TX	Fruit	420	0.01	4.2	4.2
3	Tom	Bananas	02-01-2021	TX	Fruit	120	0.01	1.2	5.4
4	Mat	Bananas	02-01-2021	NY	Fruit	30	0.01	0.3	0.3
5	Mat	Bananas	02-01-2021	NY	Fruit	50	0.01	0.3	0.6
6	Mat	Apples	03-01-2021	NY	Vegetable	80	0.02	1.6	1.6

Answered By: Pedro Rocha

Is there a way to calculate the running total across only a few columns (unique values only)?

Question:

Answers: