Hopefully you took my advice to use Text::CSV instead of a regular expression pattern to parse the UTF-8 CSV records. The following script neatly parses the example record in your original post.

#!perl

use strict;
use warnings;

use Text::CSV;

my $csv = Text::CSV->new({ auto_diag => 1, binary => 1 });

my $file = shift;

open my $fh, '<:encoding(UTF-8)', $file;

binmode STDOUT, ':encoding(UTF-8)';

my ($field1, $field2, $field3);

$csv->bind_columns(\($field1, $field2, $field3));

while ($csv->getline($fh)) {
    print "$field1\n";
    print "$field2\n";
    print "$field3\n";
}

exit 0;
[download]