In python, is there a way to vectorize adding the previous row to the current one?

Question

Is there a way for me to vectorize adding the values from one row to all the subsequent rows in a vectorized manner in python?

I was trying to use dataframe to get there but was struggling to vectorize it. This can be done with a for loop.

For example, if I use a table like this:

Table A

Index	A	B	C
1	10	11	12
2	0	0	0
3	0	1	2
4	7	0	1

and add this to another table:

Table B

Index	A	B	C
1	0	0	10
2	0	0	0
3	0	0	0
4	0	0	0

the result would be:

Index	A	B	C
1	10	11	22
2	10	11	22
3	10	12	24
4	17	12	25

The for loop of this would be something like this:


tableA.ix[0, :] = tableA.ix[0, :] + tableB.ix[i, :]

for i in range(1, len(tableA.index)):
   tableA.ix[i, :] = tableA.ix[i-1, :] + tableB.ix[i, :]

Asked By: Sleo

||

Source

Answer 1

Try this;

df3 = df1.cumsum(axis = 0).add(df2.cumsum(axis = 0), fill_value=0)
df3["INDEX"] = range(1,df3.shape[0]+1)

# Output;
   INDEX   A   B   C
0      1  10  11  22
1      2  10  11  22
2      3  10  12  24
3      4  17  12  25

Answered By: Sachin Kohli

In python, is there a way to vectorize adding the previous row to the current one?

Question:

Table A

Table B

Answers: