Data step is much faster than proc sql for large dataset, but have to sort first

It is not that bad. But the join lightly beats a subquery. The inner join is the best.

data a;
   do id=1 to 1e6;
      price=id;
          output;
          output;
        end;
run;

data b;
   do id=100000 to 200000;
          output;
        end;
run;

proc sql _method stimer;
          create table temp as
          select id,
                   sum(price) as income
         from a
         where  id in (select id from b)
                 ;
         quit;

proc sql _method stimer;
          create table temp as
          select b.id, sum(price) as income
          from b left join a on
          b.id=a.id;
quit;


proc sql _method stimer;
          create table temp as
          select b.id, sum(price) as income
          from b inner join a on
          b.id=a.id;
quit;

Learn a lot! Thank you very much!