How to label rows that don't match

[bourdeax]

Hello Everyone

First, please take a look at this data set:

ID     College1     College2     NewVariable
1            A                  A                YES
1            A                  B                YES
1            A                  A                YES
2            A                  B                 NO
2            A                  B                 NO
2            A                  B                 NO
3            A                  A               YES
3            A                  A               YES
3            A                  A               YES

My goal is to create NewVariable. Basically, if none of the values of College2 = College1, then NewVariable = No. If all the values of College2 = College1, or if some of the values of College2 = College1, then NewVariable = Yes.

Thank you!

[PaigeMiller · Accepted solution]

proc sql;
	create table want as select *,sum(college1=college2) as num
		from have group by id;
quit;

basically, this creates a new variable NUM which counts the number of matches. I will leave it up to you to turn the 0 match case into "NO" and the cases where there is 1 or more matches into "YES".

[bourdeax]

Thank you, Paige. That worked. Also, I appreciate how you answered my main question but left some work for me to do. That is a good teaching technique.

[bourdeax]

Also, in case anyone wants to know what I did after Paige's section of code:

data want;
     set have;
     if num = 0 then flag = "N";
         else flag = "Y";
run;

[Kurt_Bremser]

Data step solution:

data want;
set have;
by id;
retain newvar;
if first.id then newvar = 'N';
if college1 = college2 then newvar = 'Y';
if last.id;
keep id newvar;
run;

If needed, you can join that back to the original dataset.