Perl: Inserting an XML::Twig node with XML::Twig

Question

I am comparing two XML files. If I find a node missing in one of the files I want to insert it into the other file. Here is how I have been trying it:

my $out_file = 'fbCI_report.xml';
open my $fh_out, '>>', $out_file or die "Can't open $out_file for writing: $!";

my $currentReport = XML::Twig->new( pretty_print => 'indented' );
$currentReport->parsefile($path_to_currentReport);
print "Loaded current report.\n";

my $newReport = XML::Twig->new( pretty_print => 'indented' );
$newReport->parsefile($path_to_newReport);
print "Loaded new report.\n";

my $currentRoot   = $currentReport->root;             # get the root
my $currentBuilds = $currentRoot->first_child();      # get the builds node
my $currentXCR    = $currentBuilds->first_child();    # get the xcr node

my $newRoot   = $newReport->root;                     # get the root
my $newBuilds = $newRoot->first_child();              # get the builds node
my $newXCR    = $newBuilds->first_child();            # get the xcr node

my @currentXCRarray = $currentBuilds->children('xcr');
my @newXCRarray     = $newBuilds->children('xcr');
my $numberOfxcr     = $newBuilds->children_count();

foreach my $currentXCRmod ( @currentXCRarray ) {

    my $currentID = $currentXCRmod->att("id");

    foreach my $newXCRmod (@newXCRarray) {

        my $newID = $newXCRmod->att("id");

        if ( $newID == $currentID ) {
            last;
        }
        elsif ( $count == $numberOfxcr && $newID != $currentID ) {
            my $insert = $currentBuilds->insert_new_elt($newXCRmod);
            print "XCR does not exist in current report, adding it..\n";
        }

        $count++;
    }
}

print $fh_out $currentReport->sprint();
close $fh_out;

However this does not insert the node with the corresponding children but what I guess is the reference to the node: <XML::Twig::Elt=HASH(0x326efe0)/>. Is there a way to insert the node properly? I have yet to find anything on the CPAN site.

Sample data, current.xml:

<project>
  <builds>
    <xcr id="13367" buildable="false">
        <artifact name="rb"/>
        <artifact name="syca"/>
    </xcr>
    <xcr id="13826" buildable="false">
        <artifact name="dcs"/>
    </xcr>
  <\builds>
<\project>

new.xml:

<project>
<builds>
    <xcr id="13367" buildable="false">
        <artifact name="rb"/>
        <artifact name="syca"/>
    </xcr>
    <xcr id="13826" buildable="false">
        <artifact name="dcs"/>
    </xcr>
    <xcr id="10867" buildable="true">
        <artifact name="smth"/>
        <artifact name="top"/>
        <artifact name="tree"/>
    </xcr>
<\builds>
<\project>

http://stackoverflow.com/questions/29009370/assembling-xml-in-perl — Sobrique, Jul 21 '16 at 12:13
I'd also suggest rather than just doing `first_child` to walk your tree, doing it explicitly with `xpath` is the way to go. — Sobrique, Jul 21 '16 at 12:16
So with those XML samples, are you looking to do a merge based on `xcr` id? — Sobrique, Jul 21 '16 at 12:29
@Sobrique yes, in a way that is exactly what I want to do here. I wonder why that word never crossed my mind. — UsefulUserName, Jul 21 '16 at 12:34
Bidirectional , or just from 'new' into 'current'? And if there's a duplicate, how's that handled (if it's not the same?) — Sobrique, Jul 21 '16 at 12:38
Just from new to current. I'm not sure I understand you question about duplicates, if you are talking about ID then I am sure that they are unique. There should be only 2 cases: Either the xcr-node with the id exists, then do nothing; Or the node does not exist, then copy the missing node from new to current. — UsefulUserName, Jul 21 '16 at 12:42

score 3 · Answer 1 · answered Jul 21 '16 at 12:20

3

You should probably move the node (I can't remember what happens when you try to insert an element that's already part of a tree). So write $newXCRmo->move( first_child( $currentBuilds)) and see if this improves the situation.

I don't have much time to look at your code, so there may be other problems with it.

answered Jul 21 '16 at 12:20

mirod

15,923
3
45
65

1

The message is *"cannot paste an element that belongs to a tree"* – Borodin Jul 21 '16 at 13:29

Sobrique · Accepted Answer · 2016-07-21T12:38:13.963

You're right - that's the stringified text of an XML::Twig::Elt.

The problem is - insert_new_elt creates a new element. So what you're doing is effectively, "printing" the element id ( XML::Twig::Elt=HASH(0x326efe0)) and creating a new node called that.

But you don't want to do that - you're wanting to copy an existing one.

So I would suggest what you want to do is:

my $copied_elt = $currentXCRmod -> copy;
$copied_elt -> paste ( last_child => $currentBuilds );

Which will transfer the element (into the 'last_child' position).

Although I'd suggest that your loop is perhaps something you could improve on too - I would suggest you look at a twig_handler, to check which ID's exist in the file at parse:

my %seen_id; 
sub collect_ids {
   my ( $twig, $element ) = @_;
   $seen_id { $element->att('id') } ++; 
}

And then call this at parse time:

my $currentReport = XML::Twig->new(twig_handlers => { 'xcr' => \&collect_ids}, 
                                   pretty_print=>'indented');
$currentReport->parsefile($path_to_currentReport);

And this will let you easily compare/copy which ones do or don't exist.

Or alternatively (based on your XML sample so far):

#!/usr/bin/env perl

use strict;
use warnings 'all';

use Data::Dumper;
use XML::Twig;

my $current = XML::Twig -> new ( ) -> parsefile ('test1.xml');
my $new = XML::Twig -> new (  ) -> parsefile ( 'test2.xml'); 

my $cur_builds = $current -> root -> get_xpath('./builds',0);

foreach my $xcr ( $new -> findnodes('//xcr') ) {
   my $id = $xcr -> att('id'); 
   if ( not $current -> findnodes("//xcr[\@id=\"$id\"]") ) {
      print "$id not in current, copying\n"; 
      my $copy = $xcr -> copy; 
      $copy -> paste ( last_child => $cur_builds ); 
   }
}

$current -> set_pretty_print('indented_a');
$current -> print;

Borodin · Answer 3 · 2016-07-21T13:58:29.893

You have your comparison loop "inside out"

In addition, the test $count == $numberOfxcr will never succeed because the loop foreach my $newXCRmod (@newXCRarray) will terminate before that is true

Here's an improved version of your code that uses XPath expressions as well as any from List::Util to make the loops more concise

use strict;
use warnings 'all';

use XML::Twig;
use List::Util 'any';

my ( $path_to_curr_report, $path_to_new_report ) = qw/ current.xml  new.xml /;
my $out_file = 'fbCI_report.xml';


my $curr_report = XML::Twig->new->parsefile($path_to_curr_report);
my $new_report  = XML::Twig->new->parsefile($path_to_new_report);

my ($curr_builds) = $curr_report->findnodes('/project/builds');

for my $new_xcr_mod ( $new_report->findnodes('/project/builds/xcr') ) {

    my $new_id = $new_xcr_mod->att('id');

    next if any { $new_id eq $_->att('id') } $curr_report->findnodes('/project/builds/xcr');

    print qq{XCR with ID "$new_id" does not exist in current report. Adding it.\n};
    $new_xcr_mod->copy->paste( last_child => $curr_builds );
}

{
    $curr_report->set_pretty_print('indented');
    open my $fh, '>', $out_file or die "Can't open $out_file for writing: $!";
    $curr_report->print($fh);
    close $fh;
}

output

XCR with ID "10867" does not exist in current report. Adding it.

<project>
  <builds>
    <xcr buildable="false" id="13367">
      <artifact name="rb"/>
      <artifact name="syca"/>
    </xcr>
    <xcr buildable="false" id="13826">
      <artifact name="dcs"/>
    </xcr>
    <xcr buildable="true" id="10867">
      <artifact name="smth"/>
      <artifact name="top"/>
      <artifact name="tree"/>
    </xcr>
  </builds>
</project>

Perl: Inserting an XML::Twig node with XML::Twig

3 Answers3

output

Linked