用hash吧，相当快。




SAS博客列表http://saslist.com

According to my test, it should be reasonable fast with sql (cpu=5 secons).

286  proc sql;
287    create table mean_stat as
288    select t1.id as id ,
289           mean( x1, x2, x3, x4, x5, x6, x7, x8, x9, x10 ) as meanx,
290           mean( y1, y2, y3, y4, y5, y6, y7, y8, y9, y10 ) as meany
291    from t1 as a,
292         t2 as b
293    where t1.id=t2.id
294
295    ;
NOTE: Table WORK.MEAN_STAT created, with 2000000 rows and 3 columns.
296    quit;
NOTE: PROCEDURE SQL used (Total process time):
      real time           10.18 seconds
      cpu time            5.10 seconds

***************pgm*****;
data t1 t2(rename=(x1-x10 = y1-y10));
   array x(10) ;
   do id=1 to 2e6;
      do j=1 to dim (x);
            x[j]=ranuni(123);
          end;
          output;
        end;
   
run;

proc sql;
  create table mean_stat as
  select t1.id as id ,
         mean( x1, x2, x3, x4, x5, x6, x7, x8, x9, x10 ) as meanx,
                 mean( y1, y2, y3, y4, y5, y6, y7, y8, y9, y10 ) as meany
  from t1 as a,
       t2 as b
  where t1.id=t2.id

  ;
  quit;

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!