Compare a variable with a subset of the same variable

[jeka1212]

Hello all,

II need to compare a variable (y1) to a subset of the same variable  (y2) to generate a binary variable (beta).  is there a clever way of doing this without having missing data?

Many thanks, 

y1	y2
28	28
15.5	15.5
8.2	8.2
3.4	3.4
17.3	17.3
15.2	15.2
32.9	32.9
11.1	11.1
87.5
16.2
107.9
5.7
25.6
31.2
21.6
55.6
8.8
6.5
22.1
14.4
44.2
3.7
7.8
8.9
719
2106.667
24000
1715
3.6
521.5
1600
454
109.7

proc sql;
create table temp as
select
a.y1 as compare_y1,
b.y2 as compare_y2,
a.y1 <= b.y2 as beta
from training as a, training as b;
quit;

data training;
 input y1 y2;
     datalines; 
28	28
15.5	15.5
8.2	8.2
3.4	3.4
17.3	17.3
15.2	15.2
32.9	32.9
11.1	11.1
87.5	
16.2	
107.9	
5.7	
25.6	
31.2	
21.6	
55.6	
8.8	
6.5	
22.1	
14.4	
44.2	
3.7	
7.8	
8.9	
719	
2106.667	
24000	
1715	
3.6	
521.5	
1600	
454	
109.7
;
run;	


proc sql;
create table temp as
select 
    a.y1 as compare_y1,
    b.y2 as compare_y2,
    a.y1 <= b.y2 as beta
from training as a, training as b;
quit;

[novinosrin]

Are you after this?

data want;
set training(drop=y1 where=(y2 ne .)) ;
do n=1 to nobs;
set training(drop=y2) point=n nobs=nobs;
beta=y1 <= y2;
output;
end;
run;

[jeka1212]

Yes indeed this what I wanted . Many thanks novinosrin for your time

[ChrisNZ]

Like this?

proc sql;
create table temp as
select 
     a.y1 as compare_y1
   , b.y2 as compare_y2
   , a.y1 <= b.y2 as beta
from (select unique y1 from training where y1 ne .) as a
   , (select unique y2 from training where y2 ne .) as b;
quit;

[jeka1212]

Many thanks ChrisNZ. This is correct too