Since this may require recursion (as you say the values may be key/value
pairs themselves), you'll need something a bit more powerful than a
regex. I suggest looking into Text::Balanced, which can do what you
want.
Here's sample code to get you started. It does the first step of
extracting the outermost pairs.
#-------
use Text::Balanced qw(extract_multiple extract_bracketed);
my $str = '{key1,val1},{key2,val2},{key3,{key3.1,val3.1}},{key4,val4}';
my @pairs = extract_multiple (
$str,
[
sub { extract_bracketed($_[0],'{}') },
],
undef, # no maximum
1, # skip separator commas
);
print join "\n", @pairs;
#-------
- Mark.
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On
> Behalf Of Darryl Ross
> Sent: Wednesday, November 16, 2005 11:10 PM
> To: perl-unix-users@listserv.ActiveState.com
> Subject: [Perl-unix-users] Help Needed With a Regex
>
> Hey All,
>
> I have a string that looks like this:
>
> {key1,val1},{key2,val2},{key3,val3},{key4,val4},{key5,val5},...
>
> I'm trying to write a regular expression which can split up
> that string
> and give me a hash back. There are a few gotchas though:
>
> - The key and/or the value may or may not be inside quotes,
>
> - The key and/or the value may contain a comma (it will
> be quoted if
> it does),
>
> - There can be a varying number of key/value pairs, including only
> one pair,
>
> - The values may be key/value pairs themselves.
>
> My first attempt was something like: /({[^}]+,[^}]+})+/
>
> But that completely fails at least the last point, if not the others.
>
> Does anyone have any ideas (or pointers to good documentation)?
>
> TIA,
> Darryl
>
>
_______________________________________________
Perl-Unix-Users mailing list
Perl-Unix-Users@listserv.ActiveState.com
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
