Navigating XML file using XML::Twig

Question

I must say that I am a newbie at Perl and XML:Twig but I am a quick learner. Any help you can provide would be greatly appreciated.

Basically, I am having trouble navigating to certain nodes in an XML file so that I can extract information.

I am using a TwigHandler to get me access to a certain node in the XML, specifically the "Selection" node. The TwigHandler is working fine for me in the sense that I am able to extract some of the information I need at this level. However, there are additional nodes under "Selection" that I need to inspect and I do not know how to get to them.

I have copied a snippet of my XML below so that you can see what it looks like. In it you can see the Selection node. I am able to access the attribute "id" and the field "Name" with no problems using my Twig Handler, but I need to loop through all of those "Message" nodes under the Selection node in order to extract all of the attributes from each one of them. I have tried to get "get_xpath" to work but to no avail.

Please keep in mind that there are Message nodes under every Selection node in my XML. You only see 2 Selection nodes in the example below but in reality I could have hundreds of the "Selection" nodes with each one having "Message" nodes as children. I need to extract information from the "Message" nodes under the current "Selection" node that I am working with, i.e., I don't care about the other "Message" nodes that may be under different "Selection" Nodes.

<Selection id="54008473">
  <Name>Master</Name>
  <Contents>
    <Message refid="125796458" suppress="true" status="Unchanged"/>
    <Message refid="123991123" suppress="true" status="Unchanged"/>
    <Message refid="128054778" custom="true" status="New">
      <Content language="en"><![CDATA[<p>ada</p>]]></Content>
    </Message>
  </Contents>
  <Messages/>
  <MessagePriority>
    <Zone name="Insured Letter Intro">
      <MessageInstance id="125796375" name="LD Letter Introduction" status="Active" delivery="Mandatory" priority="1" suppressed="false" selected="true"/>
    </Zone>
    <Zone name="Insured Letter Logo">
      <MessageInstance id="125794623" name="Insured Letter Logo" status="Active" delivery="Mandatory" priority="1" suppressed="false" selected="true"/>
     </Zone>
  </MessagePriority>
</Selection>
<Selection id="54008475" datavaluerefid="54008479">
  <Name>RMBC</Name>
  <Contents>
    <Message refid="125796458" sameasparent="true" parentrefid="54008473" status="Unchanged"/>
    <Message refid="123991123" sameasparent="true" parentrefid="54008473" status="Unchanged"/>
    <Message refid="128054778" custom="true" status="New">
      <Content language="en"><![CDATA[<p>ada</p>]]></Content>
    </Message>
  </Contents>
  <Messages/>
  <MessagePriority>
     ...
  </MessagePriority>
</Selection>

Answer 1

Use findnodes() with a relative XPath in the handler for Selection to find the Contents/Message child nodes:

#!/usr/bin/perl
use warnings;
use strict;

use XML::Twig;

my %selections;

my $twig = XML::Twig->new(
    twig_handlers => {
        Selection => sub {
            #$_->print();
            print "selection id: ", $_->att('id'), "\n";

            my @messages;
            foreach my $message ($_->findnodes('./Contents/Message')) {
                #$message->print();
                print "message refid: ", $message->att('refid'), "\n";

                # store "refid" attribute in messages list
                push(@messages, $message->att('refid'));
            }

            # store collected Message nodes under selection ID
            $selections{ $_->att('id') } = \@messages;
        },
    }
);

$twig->parse(\*DATA);

while (my($id, $messages) = each %selections) {
    print "Selection '${id}' messages: @{ $messages }\n";
}

exit 0;

__DATA__
<?xml version="1.0" encoding="UTF-8"?>
<Root>
  <Selection id="54008473">
    <Name>Master</Name>
    <Contents>
      <Message refid="125796458" suppress="true" status="Unchanged"/>
      <Message refid="123991123" suppress="true" status="Unchanged"/>
      <Message refid="128054778" custom="true" status="New">
        <Content language="en"><![CDATA[<p>ada</p>]]></Content>
      </Message>
    </Contents>
    <Messages/>
    <MessagePriority>
      <Zone name="Insured Letter Intro">
        <MessageInstance id="125796375" name="LD Letter Introduction" status="Active" delivery="Mandatory" priority="1" suppressed="false" selected="true"/>
      </Zone>
      <Zone name="Insured Letter Logo">
        <MessageInstance id="125794623" name="Insured Letter Logo" status="Active" delivery="Mandatory" priority="1" suppressed="false" selected="true"/>
      </Zone>
    </MessagePriority>
  </Selection>
  <Selection id="54008475" datavaluerefid="54008479">
    <Name>RMBC</Name>
    <Contents>
      <Message refid="125796458" sameasparent="true" parentrefid="54008473" status="Unchanged"/>
      <Message refid="123991123" sameasparent="true" parentrefid="54008473" status="Unchanged"/>
      <Message refid="128054778" custom="true" status="New">
        <Content language="en"><![CDATA[<p>ada</p>]]></Content>
      </Message>
    </Contents>
    <Messages/>
    <MessagePriority>
      ...
    </MessagePriority>
  </Selection>
</Root>

Test run:

$ perl dummy.pl
selection id: 54008473
message refid: 125796458
message refid: 123991123
message refid: 128054778
selection id: 54008475
message refid: 125796458
message refid: 123991123
message refid: 128054778
Selection '54008473' messages: 125796458 123991123 128054778
Selection '54008475' messages: 125796458 123991123 128054778