Regex: Optional/alternative catpures

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

This does what I want

my $line = q{one <two>};
$line =~ /
    ([^<]+)
    (?:
        <
        (two|three)
        >
    )?
/x;

print $1, $2;
[download]

one two
[download]

The <two> is optional and that works too. But I need

my $line = q{one <four>};
[download]

to return one <four> in $1 and for $2 to be undef. Only capture to $2 if a pattern is matched otherwise capture the whole lot to $1.

Is this something that a look around assertion could handle? I've yet to conquer the blighters :-(

It will be used like

while ($line =~ /pattern/g){
  # do something with $1 and $2
}
[download]

and a typical string

one [link|file.html] two [email|me@home.com] three [not a link]
[download]

What I have at the moment is fine but I'm not picking up three [not a link] in $1

  while (
    $line =~ 
      /
        (
          [^\[]*
        )
        (?:
          \[
          (
            (?:link|email|pdf)
            [^\]]*
          )
          \]
        )?
      /xg
  )
  {
    my $txt = $1;
    my $link = $2;
    if ($link){
      my ($type, $address, $txt) = split(/\|/, $link);
      # more stuff
    }
  }
[download]

tia

Comment on Regex: Optional/alternative catpures Select or Download Code

Replies are listed 'Best First'.
Re: Regex: Optional/alternative catpures (two) by tye (Sage) on May 23, 2009 at 16:22 UTC
Minor adjustment: `$line =~ / ^ (.+?) (?: < (two\|three) > )? $ /xs;` [download] Another approach: `my $pre= $line; my $post= ''; if( $pre =~ s/ \s* < (two\|three) > $ //xs ) { $post= $1; }` [download] - tye	[reply] [d/l] [select]