SAS 9.2

data outdsn;
  set indsn;
  array zero(*) _numeric_;
  do i = 1 to dim(zero);
    if zero(i) = . then zero(i) = 0;
  end;
run;

data a;
input  a b c $;
datalines;
1 3 a
. 5 .
. 6 b
1 7 .
;
run;

data b;                    
set a;   

array chr(*) _character_;      
array num(*) _numeric_;   

do i = 1 to dim(chr);         
if  missing(chr(i)) then chr(i)='0';
end;                           
do j = 1 to dim(num);           
if  missing(num(j)) then num(j)=0;     
put _all_;
end;   
drop i j;

run;

data a;
input a b c $;
datalines;
1 . a
. 5 .
;
run;
options symbolgen mprint mlogic;
%macro missing(lib=work ,dsn=a );
proc sql noprint;
select name into :col1 - :col9999
from dictionary.columns
where libname=upcase("&LIB")
and memname=upcase("&DSN");
quit;
data _null_;
if 0 then set a nobs=nobs;
call symput('nobs',put(nobs,12.-l));
stop;
run;
data b;
set &&LIB..&DSN;
%do i=1 %to &sqlobs;
if &&col&i=" " then &&col&i=0;

%end;
run;
%mend;
%missing;

楼上的集中方法都很好，学习了~

采用array 分别对数值型和字符型变量缺失值赋值