One way to strip out the "-" characters is like this:

#!usr/bin/perl
use strict;
use warnings;

foreach my $ssn qw(123-45-6789 
                   987654321)
{
   my $digits = $ssn;
   $digits =~ s/-//g;
   print "$ssn   \t$digits\n";
}
__END__
prints:
123-45-6789     123456789
987654321       987654321
[download]

I am not sure of the best way to handle this "sometimes field 2 vs 4" without seeing a few example lines of these databases. Don't post any real SSNs!

As mentioned before, your HUGE performance gain will come by processing each of the 2 files only once. Process file 2 first to make a memory structure, then process file 1 line by line. Each file only should be read once.