XML::Rules: Can hierarchy be changed?

[hoppfrosch]

Hi, I've got a simple xml:

<root>
  <parent>p1 p2 <ch1>c1_1</ch1> p3 <ch2>c2</ch2> p4 <ch1>c1_2</ch1> p5
+</parent>
</root>

What I want to do is parsing this into a hash:

'root' => {
          'parent' => [
                      {
                        'text' => 'p1 p2 p3 p4 p5'
                        'ch2' => [
                                 'c2'
                               ],
                        'ch1' => [
                                 'c1_1',
                                 'c1_2'
                               ]
                      }
                      ]
          }

What I tried is (with XML::Rules):

my $parser =  XML::Rules->new (
  rules => [
    root => 'no content',
    parent => 'no content array',
    ch1 => 'content array',
    ch2 => 'content array',
    ]
);
my $result = $parser->parsestring($xml);

What I cannot figure out yet, is howto parse the "contents" of tag <parent> (I have to consider "p1 p2 "." p3 "." p4 "." p5" as content for some reason) into a child "text" of the item "parent" in my (desired) hash .... I tried several things like

# different Rule for "parent"
  parent => sub {$_[0]->{text} => $_[1]->{_content}},

but had no success yet ... (It's clear that above syntax does not make sense - but I think it expresses what I want to achieve)
Any help welcome!

[Jenda]

my $parser =  XML::Rules->new (
  rules => [
    root => 'no content',
    parent => sub {$_[1]->{text} = delete $_[1]->{_content}; return '@
+parent' => $_[1]},
    ch1 => 'content array',
    ch2 => 'content array',
    ]
);

The attributes are all in the hash referenced by $_[1], the $_[0] is just a string. So in the rule for <parent> you first move the text content from {_content} to {text} and then return the tag name and attributes prepending the '@' to make sure you end up with an array of parents. See the builtin rules in XML::Rules docs and especially the sub{} next to the 'as array' rule.

[murugu]

hoppfrosch,

Below code should form the required data structure as mentioned by you in the question. By the way i cant commit that my code is the efficient answer. Other revered monks will be able to give you the answers in most efficient way

#!/usr/bin/perl
use strict;
use XML::Rules;
use Data::Dumper;

my $xml = q(<root>
  <parent>p1 p2 <ch1>c1_1</ch1> p3 <ch2>c2</ch2> p4 <ch1>c1_2</ch1> p5
+</parent>
</root>);

my $parser =  XML::Rules->new (
  rules => [
    root => 'no content',
    parent => sub {    $_[1]->{text} = $_[1]->{_content}, delete $_[1]
+->{_content}, return ($_[0] => [$_[1]] , [$_[0] => $_[1]])} , 
    ch1 => 'content array',
    ch2 => 'content array',
    ]
);
my $result = $parser->parsestring($xml);


print Dumper $result;

[Jenda]

sub { $_[1]->{text} = $_[1]->{_content}, delete $_[1]->{_content}, ret
+urn ($_[0] => [$_[1]] , [$_[0] => $_[1]])}

is better written as

sub { $_[1]->{text} = $_[1]->{_content}; delete $_[1]->{_content}; ret
+urn ($_[0] => [$_[1]] , [$_[0] => $_[1]])}

You can put several statements into an anonymous subroutine so no need to use commas where you mean semicolons :-)