Pandas cumsum with keys-CodePudding

I have two DataFrames (first, second):

index_first	value_1	value_2
0	100	1
1	200	2
2	300	3

index_second	value_1	value_2
0	50	10
1	100	20
2	150	30

Next I concat the two DataFrames with keys:

z = pd.concat([first, second],keys=['x','y'])

My goal is to calculate the cumulative sum of value_1 and value_2 in z considering the keys. So the final DataFrame should look like this:

index_z	value_1	value_2
x,0	100	1
x,1	300	3
x,2	600	6
y,0	50	10
y,1	150	30
y,2	300	60

CodePudding user response：

Use GroupBy.cumsum by first level created by keys from concat:

df = z.groupby(level=0).cumsum()
print (df)
               value_1  value_2
  index_first                  
x 0                100        1
  1                300        3
  2                600        6
y 0                 50       10
  1                150       30
  2                300       60