Yes. But I still thank the proc freq is better. It is designed for it.
517 data t1;
518 do i=1 to 10;
519 n=ceil(ranuni(123)*10000000);
520 do j=1 to n;
521 output;
522 end;
523 end;
524 keep i;
525 run;
NOTE: The data set WORK.T1 has 42308340 observations and 1 variables.
NOTE: DATA statement used (Total process time):
real time 2.48 seconds
cpu time 2.02 seconds
526
527
528
529 proc means data=t1(keep=i) nway noprint;
530 class i;
531 output out=freq;
532 run;
NOTE: There were 42308340 observations read from the data set WORK.T1.
NOTE: The data set WORK.FREQ has 10 observations and 3 variables.
NOTE: PROCEDURE MEANS used (Total process time):
real time 2.46 seconds
cpu time 6.30 seconds
NOTE: There were 42308340 observations read from the data set WORK.T1.
NOTE: The data set WORK.FREQ2 has 10 observations and 3 variables.
NOTE: PROCEDURE FREQ used (Total process time):
real time 3.22 seconds
cpu time 2.82 seconds