Compress and leading zeros if needed in datastep

[Q1983]

data have;
length ssn $9;
input ssn;
datalines;
000000000
112334588
426-277
;run; 

Creates the following output

ssn
000000000
112334588
426-277

I need to do the following;

if the ssn is a blank or if there is any character except consecutive 9 numbers I need it to show as 000000000  where this is a character

Does compress work in this situation?  I have not really used compress much.  So in the example here I need for 426-277 to read 000426277

[mkeintz]

Instead of using compress to remove characters, you can use it to keep characters, as in:

ssn=compress(ssn,'0123456789','k');

where 'k' says to keep the characters in the 2nd argument.

You can then prefix the resulting ssn with 9 zeroes, and take a substr of the last 9 characters:

data have;
  length ssn $9;
  input ssn $9.;
datalines;
000000000
112334588
426-277  
            
run; 

data want;
  set have;
  put 'Before: ' ssn= @;
  ssn=compress(ssn,'0123456789','k');
  put @25 'between: ' ssn= @ ;
  ssn=substr(cats('000000000',ssn),lengthn(ssn)+1);  
  put @45 '  After:' ssn  ;
run;

The substr function tells sas to take characters starting at position lengthn(ssn)+1.   lengthn will return a 0 for a blank argument, while length will return a 1).   I've added a fourth blank SSN to the sample data. 

This means for ssn='426277' (after the first compress), you would get

• 000000000426277  as the first arg of the 2nd compress
• 7  as the 2nd arg of the 2nd compress

which means the substr starts in position 7 of 000000000426277, yielding   000426277.