干儿子还是不行的，还是要靠亲儿子啊
个人感觉是data step里稍微方便点

嗯..目前只想到用first.分组变量来搞了...

Here are several ways to group your data.

data t1;
   call streaminit(123);
   do i=1 to 100;
      x=rand('chisquare',2);
          output;
        end;
run;


proc sql;
  select x, int((i-1)/10) as group
  from t1
  order by 2
  ;
quit;

proc sql;
  select x, int((monotonic()-1)/10)  as group
  from t1
  order by 2
  ;
quit;

proc sql;
  select x, ceil(10*ranuni(123)) as group
  from t1
  order by 2
  ;
quit;