How can I parse CSV data, assuming that the string is encoded in Shift-JIS?

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

I have tried the parse_csv routine from the Perl Cookbook, and it works fine on ascii encoded strings, but as soon as I start using shift-jis, which is a multi-byte character encoding, I end up with wierd results, because sometimes it matches in the middle of a character.

Right now I am using a real parser working character by character, but it is really slow... The problem comes from this line, I think:

my $sjis = q{
             [\x00-\x7F]
            |[\x81-\x9F\xE0-\xFC][\x40-\x7E\x80-\xFC]
            |[\xA0-\xDF]
           };

@chars = $text =~ /$sjis/gox;
[download]

Any ideas on how to speed this/avoid backtracking?

Thanks

Originally posted as a Categorized Question.

Comment on How can I parse CSV data, assuming that the string is encoded in Shift-JIS? Download Code

Replies are listed 'Best First'.
Re: How can I parse CSV data, assuming that the string is encoded in Shift-JIS? by AgentM (Curate) on Oct 18, 2000 at 05:06 UTC
This may involve more programming than you're interested in, but you can use Jcode at CPAN to come up with your own multibyte char techniques. It's very nice. It might improve your parser to a point where the regexes look more logical.	[reply]