Good follow-up, and here's a snippet:
Just FYI, there are multiple CSV parsing modules on the CPAN. I use Text::CSV_XS personally.
sub parse_line {
my $line = shift;
chomp($line);
print "LINE: $line<br>\n" if $debug;
my %record;
my $entry;
my $i = 1; # First index
while ($line) {
if ($line =~
s {
^\"
((?:[^\"]|\"\")*)
\"
(?:,|$)
} {}x) {
" is not a special character, in a regular expression, so save your eyes and drop the \s. ;)
$entry = $1;
} elsif ($line =~
s {
^
(.*?)
(?:,|$)
} {}x) {
$entry = $1;
} else {
die "Can't parse the line $line";
}
$entry =~ s/\"\"/\"/g;
This line is out of place, isn't it? It's only needed if the field was quoted and should be moved inside the if.
$record{$i++} = $entry;
You can eliminate the need for this line and the entry variable, if you store them when you find them.
Also, you're using a hash when you should be using an array. Numerically indexed data belongs in an array.
} return \%record; }
Unfortunately, this sub doesn't really tell us about your problem. I see nothing wrong here. Does %record contain what you think it does on exit? You might try printing it to find out.
I suspect the original problem is in your output code somewhere.
James
-- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] <http://learn.perl.org/> <http://learn.perl.org/first-response>
