CODEWITHSUNDEEP
pythonpandas
Macterror
Macterror

Reputation: 431

Pandas rolling sum on multiple columns

I have a pandas dataframe configured in the following structure:

   date  amount  another_amount  name
0   1/1       5               6  dave
1   2/1       3               4  dave
...
2  12/1       7              10  dave

How can I sum the previous row value for amount and another_amount until the last row will have the sum of the entire dataframe?

So for example the 2/1 row will have 8 (5+3) for amount and 10 (6+4) for another_amount, then the 3/1 row would have whatever that row contained plus the previous sum of 8 and 10 in their respective columns

I apologize if this a confusing question I will try to clarify as much as possible

Thanks!

Upvotes: 0

Views: 1165

Answers (1)

Dani Mesejo
Dani Mesejo

Reputation: 61910

You could use cumsum:

df.iloc[:, 1:-1] = df.iloc[:, 1:-1].cumsum()

print(df)

Output

   date  amount  another_amount  name
0   1/1       5               6  dave
1   2/1       8              10  dave
2  12/1      15              20  dave

Or if you only want amount and another_amount:

df.loc[:, ['amount', 'another_amount']] = df.loc[:, ['amount', 'another_amount']].cumsum()

Upvotes: 2

Related Questions