Davorg has good advice here. I asked the very same question a while back. Podmaster beat mirod to the punch plugging XML::Twig, which I am using in production today :-) That node also has a simple SAX version (that I'm still not quite comfortable with).

It should be noted that the first two solutions above are brittle, and will break on things such as nested <ArgusFlowRecord> tags in a CDATA section.

So, take davorg's advice. Use a real parser. But if you don't, here's a little snippet that I use when speed is of the essence. It's probably broken in some way that I haven't realized yet (and some that I have). It expects its input from STDIN.

#!/usr/bin/perl

use warnings;
use strict;
$|++;

use XML::LibXML;

my $parser = XML::LibXML->new();
my $closing_root_tag = '</ArgusFlowRecord>';
my $skip_past        = '<ArgusDataStream>';

{

  $XML::LibXML::skipXMLDeclaration = 1;

  local $/ = $skip_past;
  <>; # ignore declaration and stream open tag

  local $/ = $closing_root_tag;
  my $temp_chunk;

  while ( <> ) {
    $temp_chunk .= $_;
    my $dom;
    eval { $dom = $parser->parse_string( $temp_chunk ) };
    next if $@; # keep nibbling if XML is invalid
    undef $temp_chunk;
    print $dom->toString(), "\n"; # or do other processing
  }

}
[download]

-- dug