XML::Simple returns a hash, which has no particular order. There are multiple ways around this, if order is important to you.

It is not enough to use ForceArray in this case, because the XML that the OP presents uses key attributes. The regular elements will be forced into an array structure but the key attributes would still live in an unsorted hash.

The solution is to combine ForceArray with the option KeyAttr to also force key attributes into an array.

my $config = XMLin("meep.xlf", ForceArray => 1, KeyAttr =>[]);
[download]

NB: This changes the structure that XML::Simple returns so you will need to update your code a little.

I tend to use Data::Dumper to inspect the hash that XML::Simple returns to check if that meets my needs.

++varian:

...and that's what I get for guessing instead of trying it out.

Thanks for the catch! (Yarch, two slipped gears on the same day. I'll have to be sure not to skip my morning coffee ... um, err, uh ....oil! Yeah, that's the ticket.....)

...roboticus

Sorry for the late replies, work got over-busy and this project set on the backburner, but i'm back on it and a good step ahead thanks to your input.

And i have to say, as for quickly solving my problem with the least effort, varian is right on spot! :)
I should've known that i should've read the documentation more closely. ^^

Also, yes, i've been basically using Data::Dumper all the time, but cut it out of my example code as not relevant. I however use it to basically reverse construct my access code. =)

I suspect it's because of the hash that you're getting back, what with hashes not having an order and all...

Perhaps you can get a list of attribute names (ordered) and then get the values from them when processing the xml?

Update roboticus speaks great wisdom on this in Re: sort order of imported xml data?. It's clear he used those extra 11 minutes to figure out what he's talking about.

What about using a different module?

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

use XML::Rules;

my $parser = XML::Rules->new(
    start_rules => [
        file => sub {
            my ($tag_name, $attrs, $context, $parent_data, $parser) = 
+@_;
            $parser->{pad}{file} = $attrs->{original};
            $parser->{pad}{file} =~ s{game/stringtable/}{}i;
        },
    ],
    rules => [
        source => 'content',
        target => 'content',
        _default => '',
        'trans-unit' => sub {
            my ($tag_name, $attrs, $context, $parent_data, $parser) = 
+@_;
            print EXTR qq{"$parser->{pad}{file}","$attrs->{id}","$attr
+s->{source}","$attrs->{target}"\n};
            return;
        },
    ],
);

open(EXTR, ">meep.csv") or die $!;
$parser->parsefile( "meep.xlf");
[download]

What the script does is that when the parser parses the opening tag of <file> the script tweaks and remembers the original attribute in a "pad" - an attribute of the parser object designated to hold the script specific data. Then whenever it parses the complete <source> or <target> tag it remembers just the content and makes it readily available in the atribute hash of the parent tag and then whenever it parses the complete <trans-unit> tag it prints the remembered file name, the id attribute and the contents of the <source> and <target> subtags. And forgets the data of that tag.

Here is a version without using the pad and using a lexical filehandle:

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

use XML::Rules;

my $parser = XML::Rules->new(
    start_rules => [
        file => sub {
            my ($tag_name, $attrs) = @_;
            $attrs->{original} =~ s{game/stringtable/}{}i;
            return 1;
        },
    ],
    rules => [
        'trans-unit' => sub {
            my ($tag_name, $attrs, $context, $parent_data, $parser) = 
+@_;
            my $file = $parent_data->[-2]{original};
            print {$parser->{parameters}{FH}} qq{"$file","$attrs->{id}
+","$attrs->{source}","$attrs->{target}"\n};
            # or
            # my $FH = $parser->{parameters}{FH};
            #print $FH qq{"$file","$attrs->{id}","$attrs->{source}","$
+attrs->{target}"\n};
            return;
        },
        source => 'content',
        target => 'content',
        _default => '',
    ],
);

open(my $EXTR, ">meep.csv") or die $!;
$parser->parsefile( "meep.xlf", {FH => $EXTR});
close $EXTR;
[download]