how to substract between (only 1)corresponding variable across datasets

[kashili]

I have 2 datasets with structure <Id Year month result>. Say data is

SET1:

0  2001   01  12

1  2001   02  10

2  2001   03  13

AND set2

0  2001   01  14

1  2001   02  15

2  2001   03  16

i want only the "result" of set 2 to be substracted from the corresponding rows of set1 (where id year and month match)..result set would be

0  2001   01  -2

1  2001   02  -5

2  2001   03  -3

I use proccompare, but it diffs both the datasets across all the corresponding elements.  How can I diff only the corresponding "results" and preserve "id year month"..thank you

[Linlin · Accepted solution]

try:

data want;

merge set1(in=a rename =(result=r1)) set2(in=b rename=(result=r2));

by id year month;

if a and b;

r_diff=r1-r2;

run;

[kashili]

thanks Linlin