I typically write this type of code using PROC SQL statements but I'm trying to see if there are other ways of doing it using more inherent SAS code.
I have a dataset that I want to summarize that has multiple recods per the class column that I'll be summarizing by. What I would like to do is conditionally summarize the data (sales) based on another column in the set.
Here is what I have:
classVar: what I will summarize by (the CLASS statement)
numVar1, numVar2 numVar3: the columns for the VAR statement
txtVar1: the column that I want to conditionally summarize by
This works to get totals (I know there are other nuances of writing the code):
proc summary data=mydata;
var numVar1 numVar2 numVar3;
types classVar; /* do not include the "total" row */
Does anyone know if you can conditionally sum .. for example if txtVAr1 has two different options ("Yes" and "No" for example), can proc summary actually generate 6 output columns instead of just the 3 as seen above?
I've google'd this to no result so if anyone has any insight, it would be greatly appreciated. So far I'm coming to the conclusion that it's not possible and I just need to go back to PROC SQL and write the scripts myself.
1) create additional "count" columns based on some variable condition, or
2) split your file into two files, perform two PROC SUMMARY executions under difference rqmts and then bring the two files back together (and maybe requiring a third PROC SUMMARY).