With small files (say, < 100MB) , I usually just load the whole thing in, and then get all the elements I'm interested in, and work on them - I find it easier than the stream interface. However, in your case, we might be stretching it unless you've got a 64-bit perl. So here's a stream example. Note that for your example, I just put the whole thing in "<something>..</something>" tags so it'd be valid XML.

use strict;
use warnings;
use XML::Twig;

my $twig = XML::Twig->new(
                          twig_handlers => {
                              offer => sub {
                                  printf("%s,%s,%s\n",
                                         $_->att('id'),
                                         $_->first_child('f1')->text()
+,
                                         $_->first_child('f2')->text()
+,
                                        );
                              }
                          }
                         );
$twig->parsefile('a.xml');
[download]

Hope that helps

Why did you post this when you already have a thread on the same topic? Besides, show some effort, read document, try it. Come here again, if you runinto trouble.

You have no respect to other people's effort.

Some of the posts use Katrina's name are not real, and even it is real, lots of those projects are just frauds. I don't want to be used in the wrong way.

I am sorry, if that is not the way you do things here. I am new at this please give me a break.

Whether or not the invocation of Katrina is real or designed to solicit free help doesn't matter much, IMHO. Take a look at some of the real suggestions offered here for help on the question of parsing large XML files. :-)