proc transpose data=a out=b(rename=(_name_=var));
var _numeric_;
run;
data wanted;
set b;
count=sum(of col:);
keep var count;
run;

It works! Thanks!怎么发钱呢？

Thx,学习了一个反转 proc transpose.

First doing the sum and second transposing the results. This will be more efficient.

You can do it within ONE data step.


data t1;
array x(*) a b c d e f g h j k;
do _i_=1 to  20;
   do _j_=1 to dim(x);
     x(_j_)=rannor(123)>0;
   end;
output;
end;
keep a b c d e f g h j k;
run;

data t2;
  set t1 end=end;
  array x(*) a b c d e f g h j k;
  array sx(*) sa sb sc sd se sf sg sh sj sk;
  retain sa sb sc sd se sf sg sh sj sk;
   do _j_=1 to dim(x);
     sx(_j_)+x(_j_);
   end;
   if end then do;
    do _j_=1 to dim(x);
     name=vname(x(_j_));
         count=sx(_j_);
         output;
   end;
   end;
   keep name count;
run;

proc print;run;



proc means data=t1 noprint;
var a b c d e f g h j k;
output out=sum sum=;
run;

proc transpose data=sum out=t3(rename=(_name_=name col1=count));
var a b c d e f g h j k;
run;

proc print;run;