你去搜一下
论坛里面曾经有一些人写过这些代码

data crackman;
do date=1 to 1000;
x=date;
output;
end;
run;
proc transpose data=crackman out=a;
run;
%macro mean();
%do i=1 %to 941;
%let j=%eval(&i.+59);
s&j.=mean(of col&i.-col&j.);
%end;
%mend mean;
data result;
set a;
array s s60-s1000;
%mean();
keep s:;
run;
proc transpose data=result out=result(keep=col1);
var _all_;
run;
data crackman;
set crackman;
obs=_n_;
run;
data result;
set result;
obs=_n_+59;
run;
proc sql;
create table b as select crackman.date,crackman.x,result.col1 from crackman join result on crackman.obs=result.obs;
quit;


写了一个
可以实现
但是效率肯定不高

3# crackman
还是非常感谢

proc means就可以啊，很简单

搞错了，呵呵