XML parsing help need, please...

[spstansbury]

I am trying to extract some information from xml report generated by a trouble ticket system.

I have an "incident" number, and want to extract additional information from that node.

The script looks like this:

#!/usr/bin/perl

use strict;
use warnings;

use XML::LibXML;
use XML::LibXML::XPathContext;

my $file =  "./incident.xml";

my $parser = XML::LibXML->new(XML_LIBXML_RECOVER => 2);
   $parser->recover_silently(1);

my $dom    = $parser->load_xml( location => $file );
my $root   = $dom->getDocumentElement();
my $xpc    = XML::LibXML::XPathContext->new($root);

##############################################
# If we have an incident ticket reference get info from that ticket.
    
my $inc_node = $xpc->findnodes('/upload/incident[number/text() = "INC9
+02531"]');

my $service_restoration_time = $xpc->findvalue( "./u_service_restorati
+on_time", $inc_node );
my $assignment_group = $xpc->findvalue( ".//assignment_group/\@display
+_value", $inc_node );

# etc...

print $service_restoration_time . "\n" ;
print $assignment_group . "\n" ;

The XML file that the script parses looks like this (shortened, has thousands of incident entries. There is no namespace declared, not sure how to handle that...

I'm getting "XPathContext: lost current node at ./query_test.pl line 23" as the error.

<?xml version="1.0" encoding="UTF-8"?>
<unload unload_date="2010-03-17 13:02:24">
<incident action="INSERT_OR_UPDATE">
<opened_at>2009-12-15 14:29:55</opened_at>
<closed_at>2009-12-21 05:00:21</closed_at>
<u_problem_summary/>
<u_item>none</u_item>
<business_duration>1970-01-03 05:30:05</business_duration>
<category>none</category>
<number>INC902550</number>
<u_caller_location/>
<u_service_restoration_time>2009-12-15 23:01:17</u_service_restoration
+_time>
<u_close_code>Service Restored</u_close_code>
<u_service_loss_time>2009-12-15 14:29:55</u_service_loss_time>
<u_user_issue>Software Issue/Bug</u_user_issue>
<element_id>92becbd70a0a140b0187537b8516f18c</element_id>
</unload>

Would greatly appreciate being pointed in the right direction!

[ikegami]

findnodes (note the "s") returns a list of nodes (list context) or a NodeList object (scalar context). You want an actual node. If you want to discard all but the first result, you can use a list assignment:

   |         |
   v         v
my ($inc_node) = $xpc->findnodes('/upload/incident[number/text()="INC902531"]');

But your XPath is wrong. It should be

                                     |                                 |
                                     v                                 v
my ($inc_node) = $xpc->findnodes('/unload/incident[number/text()="INC902550"]');

[spstansbury]

Discarding all but the first works fine, as there should only be one match. Thank you for your help!

[ikegami]

There is no namespace declared, not sure how to handle that...

So the elements belong to the null namespace, which is the one XPath engines (should*) look into for a nodetest with no prefix.

If all the elements are in the null namespace, you may use XPC as you used it

my $xpc = XML::LibXML::XPathContext->new($root);
$xpc->findnodes(...)
$xpc->findnodes(..., $topic)

or you may avoid using XPC completely.

$root->findnodes(...)
$root->findnodes(..., $topic)

* — In practice, many XPath engine get this wrong. LibXML gets it right.

[spstansbury]

Thanks for pointing out the option - I didn't realize that not using XPC would work

[amedico]

In the sample XML, the incident tag isn't closed - could be a problem if it's also the case in the real input (and not just a copy&paste error here).

[spstansbury]

The latter - my apologies for the sloppy copying for the example