Your code has a pretty fundemental flaw where it assumes that all commas in the data are there to separate fields. CSV files allow commas to exist in quoted fields. That's why Text::CSV_XS (or even Text::ParseWords) are far better solutions.

But, if I knew that this wasn't going to be an issue, I'd do it something like this:

$_ = <DATA>;
chomp;
my @hdrs = split /,/;

my @recs;
while (<DATA>) {
  chomp;
  my %rec;
  @rec{@hdrs} = split /,/;
  push @recs, \%rec;
}
[download]

That seems, to me, to create a far more useful data structure - an array of records, with each record being a reference to a hash where the keys are the header fields and the values are the associated data values.

Here's an example that then sorts the data on the second value in each record:

print "num1\tnum2\tnum3\n";
foreach (sort { $a->{num2} <=> $b->{num2} } @recs) {
  print "$_->{num1}\t$_->{num2}\t$_->{num3}\n";
}
[download]

"The first rule of Perl club is you don't talk about Perl club."

As I said, the problem was probably me and not Text::CSV_XS. But so as not to be elusive I'll fill you in as to why I gave up on it.

Firstly I was working on something that had to done as fast as humanly (or inhumanly as the case sometimes is) possible, so when I encountered the problems with Text::CSV_XS I didn't spend too much time on figuring out what wrong.
I wasn't doing anything particularly fancy or neat, just formatting some data from a TDF. But for some reason it would only find 4 of 9 records, and I can't see why. There wasn't anything too gruesome in the data (as it was created to be as friendly as possible) and splitting it on "\t" seemed to work a treat so I stuck with that.

If you're interested in the code it was something like this -

use strict;
use Text::CSV_XS;

my $tdf = Text::CSV_XS->new({'sep_char'=>"\t"});
open(TDF, "somedata.tdf") or die("b0rk - $!");
my @fields = ();
while($tdf->parse(scalar <TDF>)) {
    @fields = $tdf->fields();
    print "@fields\n";
}
[download]

That seems to work fine on some simple data, but it broke on the data I was feeding it, so I'm guessing that the data was the root of the problem (and I would've dug further but I was under a *wee* bit of pressure ;o). I should imagine I'll use it if the need ever arises again, and hopefully the data behaves itself this time (assuming it was the data ...)
HTH

broquaint