Re: XML::Simple in Xerces (was Re: Xerces-perl for Win32)

Brian Faull <[email protected]> Tue, 11 Nov 2003 23:29:26 -0500

Newsgroups	gmane.text.xml.xerces-p.devel
Organization	The MITRE Corporation
Message-ID	<[email protected]>

"Jason E. Stewart" wrote:

> > Wow, thanks!
> 
> Welcome. New users always inspire me to add features to Xerces.

Glad I could help! :)

> looks like you hit the 'send' key before the 'attach' key, we didn't
> get anything.

Drat... I do that far more often than I'd like to admit... I'm typing this
*after* I attached the file this time. Sorry for the spam.

> > p.s. sorry that the subject in this thread has strayed so far from the
> > original 'subject'...
> 
> time for a subject change, I think...

Good subject! :)
-brian



"Jason E. Stewart" wrote:
> 
> Brian Faull <[email protected]> writes:
> 
> > Wow, thanks!
> 
> Welcome. New users always inspire me to add features to Xerces.
> 
> > Attached, if anyone is interested, is a modification of node2hash that
> > creates a data structure that is identical to that which XML::Simple
> > creates in XMLin(). It's not pretty, but neither is the data structure
> > used by XML::Simple. :)
> 
> looks like you hit the 'send' key before the 'attach' key, we didn't
> get anything.
> 
> > I've written one the other way, too, but it's still flaky; I'll post both
> > again when I've cleaned them up.
> 
> Ok, sounds good.
> 
> > Thanks again,
> > -brian
> 
> you're welcome.
> 
> > p.s. sorry that the subject in this thread has strayed so far from the
> > original 'subject'...
> 
> time for a subject change, I think...

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

node2hash.pl (application/x-perl, 2.6 KB)

# Construct the ridiculous data structure used by XML::Simple
# Each node (except toplevel!) is an arrayref.
# For text nodes, the hashref in the first element of the arrayref includes a key
#   'content' whose value is an arrayref of all of the text nodes within parent node.
# For attributes of a node, the first element of the arrayref includes a key having
#   the same name as the attribute name, and the value is an arrayref. If there exists
#   one or more child nodes with the same name as an attribute name, the first element
#   of the arrayref is the value of the attribute, and the remaining elements are
#   hashrefs containing one key 'content' whose value is the enclosed text.
# For children of the node, there is a key that points to an arrayref, as above.
# If there is more than one node of the same type, these are represented as individual
#   hashrefs within the enclosing arrayref.
# This is perhaps an oversimplification. Further information in XML::Simple

sub node2hash {
  my $node = shift;
  my $return = [{}];

  # build a hashref that represents the attributes of this element
  my $nodename = $node->getNodeName();
  if ($node->hasAttributes()) {
    my %attrs = $node->getAttributes();
    foreach my $attr (keys %attrs) {
      $return->[0]->{$attr} = $attrs{$attr};
    }
  }

  # recursively create structures for child nodes...
  # sorry, this gets kinda ugly...
  if ($node->hasChildNodes()) {
    my @text = ();
    foreach my $child ($node->getChildNodes()) {
      my $childname = $child->getNodeName();
      if ($child->isa('XML::Xerces::DOMElement')) {
        if (exists $return->[0]->{$childname}) {
          if (UNIVERSAL::isa($return->[0]->{$childname}, 'ARRAY')) {
            push @{$return->[0]->{$childname}}, node2hash($child)->[0];
          }
          else {
            $return->[0]->{$childname} = [$return->[0]->{$childname}, node2hash($child)->[0]];
          }
        }
        else { $return->[0]->{$childname} = node2hash($child); }
      }
      elsif ($child->isa('XML::Xerces::DOMText')) {
        my $val = $child->getNodeValue();
        push @text, $val
          if $val =~ m/\S+/;
      }
    }

    # Add text nodes as key 'content'
    if (exists $return->[0]->{'content'}) {
      if (UNIVERSAL::isa($return->[0]->{'content'}, 'ARRAY')) {
        unshift @{$return->[0]->{'content'}}, @text;
      }
      else { warn "content element should be array"; }
    }
    else {
      if (scalar @text == 1) { $return->[0]->{'content'} = $text[0]; }
      if (scalar @text >  1) { $return->[0]->{'content'} = \@text; }
    }
  }
  return $return;
}