I have seen this issue addressed in other posts and I know that a statement like: put(v1,z5.); requires that v1 be numeric. I have multiple datasets that I am trying to append where sometime v1 is numeric and sometimes is character--v1 is a zipcode and when it is numeric the leading zero sometimes needs to be restored. In files where the zipcode is charachter, the lengths vary from $5 to $10. So I want to standardize all zipcodes before appending. I have tried 2 approaches using the TYPE column from a PROC CONTENTS output to determine how to treat the field in different files, so I can build 1 macro to handle both types of cases. Approach 1: data t; input @1 zip $3. @5 typ 8.; cards; 949 2 31 2 768 2 74 2 270 2 173 2 100 2 755 2 476 2 454 2 415 2 ; data t_; set t; if typ=2 then zipcode=put(input(zip,8.),z5.); else if typ=1 then zipcode=put(z,z5.); run; Approach 2: data t; input @1 zip $3. @5 typ 8.; cards; 949 2 31 2 768 2 74 2 270 2 173 2 100 2 755 2 476 2 454 2 415 2 ; data t_; set t; if typ=2 then goto zchar; else goto znum; zchar: zipcode=put(input(zip,8.),z5.); goto continue; znum: zipcode=put(zip,z5.); goto continue; continue: x1=typ*9; x2=typ/x1; x3=(.985*typ)+(.375*x2); run; Both approaches fail to pass over the if-then-else condition that does not apply. Any suggestions? Thanks.
... View more
