Depends upon how similar the files are and exactly how you have to import them. Yes, it MAY be doable in one datastep. Below, I create two csv files, one space delimited file and one tab delimited file, and then import all 4 in one datastep. However, without knowing what SAS skills you have, or what your files actually look like, this may be an extreme oversimplification:
/*create test data*/
data _null_;
input x y z;
file "c:\art\test\csv1.csv";
put x "," y "," z;
cards;
1 2 3
4 5 6
;
data _null_;
input x y z;
file "c:\art\test\csv2.csv";
put x "," y "," z;
cards;
7 8 9
10 11 12
;
data _null_;
input x y z;
file "c:\art\test\txt1.txt";
put x y z;
cards;
13 14 15
16 17 18
;
data _null_;
input x y z;
file "c:\art\test\txt2.txt";
put x "09"x y "09"x z;
cards;
19 20 21
22 23 24
;
/*get file names*/
filename indata pipe "dir c:\art\test\*.* /b";
/*import all 4 files in 1 datastep*/
data want;
length fil2read $100;
informat fn $50.;
infile indata truncover;
input fn &;
fil2read="c:\art\test\"||fn;
if scan(fil2read,-1) eq "csv" then delim=",";
else delim=" ";
done=0;
infile dummy filevar=fil2read dlm=delim end=done;
do while(not done);
input Var1-var3;
end;
run;
