Just for interest, here's the same approach in Perl 6:

  use v6;

  my regex fields { \" <( .*? )> \" | <-[,"]>* }

  my $input  = open  'input.csv', :r;
  my $output = open 'output.tff', :w;

  for $input.lines {
      next unless /<fields>* % ','/;
      $output.print: @<fields>.join(chr 31) ~ chr 30;
  }

  $output.close;
[download]

That's not the shortest way to do it in Perl 6, but it's closest to the Perl 5 example above.

A shorter and more Perl6-ish version might be:

  my regex fields { \" <( .*? )> \" | <-[,"]>* }

  lines open 'input.csv'
    ==> map({@<fields>.join(chr 31) ~ chr 30 if /<fields>* % ','/})
    ==> spurt('output.tff');
[download]

Damian