open 935133.xml ;
for /excerpt/unit/products/article/quantity echo (../../../Name) (../a
+rt_code) (.) ;
[download]

Unfortunately XML::XSH2 seems not to be available for ActivePerl v.5.12.

You could try and install it using cpan XML::XSH2 from the command line. See also Installing Modules from the tutorials section.

Sorry choroba I forgot to say thank you, this is not my style :-)

Would you like the data structure like this?

{
  'Entity B' => {
                'Boss' => 'Name',
                'unitnumber' => '2',
                'contactinfo' => {
                                 'URL' => undef,
                                 'email' => undef,
                                 'Telefon' => {
                                              'telnumber' => '456',
                                              'directcall' => '78910',
                                              'code' => '0999'
                                            },
                                 'Address' => {
                                              'zip' => '11111',
                                              'Street' => 'SomeOtherSt
+reet',
                                              'Ort' => 'City',
                                              'Building' => '2'
                                            },
                                 'Fax' => {
                                          'telnumber' => '456',
                                          'directcall' => '10987',
                                          'code' => '0999'
                                        }
                               },
                'products' => {
                              'E5722' => 'few',
                              'C8099' => '522',
                              'F3596' => 'few',
                              'B0765' => '988',
                              'A1136' => '1982',
                              'D3938' => 'few'
                            }
              },
  'Entity A' => {
                'Boss' => 'Name',
                'unitnumber' => '1',
                'contactinfo' => {
...
[download]

use strict;
use XML::Rules;
use Data::Dumper;

my $parser = XML::Rules->new(
    rules => {
        'excerpt' => 'pass no content',
      'Address,Fax,Telefon,contactinfo,products' => 'no content',
      'Boss,Building,Name,Ort,Street,URL,art_code,code,directcall,emai
+l,quantity,quantity_small,telnumber,unitnumber,zip' => 'content',
      'article' => sub {
        if (exists $_[1]->{quantity_small}) {
            return #'%article' =>{
                $_[1]->{art_code} => 'few' # };
        } else {
            return #'%article' => {
                $_[1]->{art_code} => $_[1]->{quantity} #};
        }
      },
      'unit' => 'no content by Name',
    }
);

my $data = $parser->parse(\*DATA);

print Dumper($data);

__DATA__
<excerpt>
<unit>
<unitnumber>1</unitnumber>
...
[download]

The base set of rules was generated by: perl -MData::Dumper -MXML::Rules -e "print Dumper(XML::Rules::inferRulesFromExample( 'c:\temp\excerpt.xml'))"

Thank you Jenda,
I must read this carefully und try it. I did not know XML::Rules bevor. The good news - this module exists for ActivePerl.
The "real" huge xml file consists of many nested structures like in the example. They "dive" from the surface of simple data such as the "address" or the "boss name" (and the "unit_id").
I had hence an idea to make several csv files with id of the unit (here in example shown as unit name) and connect them in the database later. This eclectic (promiscuous? :-)) idea comes since my knowledge of perl is limited and I have to get the things run at the same time.

If the file is huge you can process it in parts. In this case and with XML::Rules it would mean that the rule for <unit> would be a subroutine that inserts the data of the unit to database and then returns nothing. That way you do not keep the already processed data in memory.

Another good module for processing huge XML files is XML::Twig.

Jenda
Enoch was right!
Enjoy the last years of Rome.

use strict;
use warnings;
use XML::LibXML;

my $filename = "Test.xml";

my $my_object = XML::LibXML->new();
my $treeobjekt   = $my_object->parse_file($filename);
my $root = $treeobjekt->getDocumentElement;

    my @articles = $treeobjekt->findnodes('//article');
    for(my $j=0;$j<@articles;$j++) {
        my $unitname = $articles[$j]->parentNode->parentNode->findvalu
+e('./Name/text()');
        my $article = $articles[$j]->findvalue('./art_code/text()');
        my $amount = $articles[$j]->findvalue('./quantity/text()')//"0
+";
    print "$unitname;$article;$amount\n";
}
[download]

Entity A;A3236;554
Entity A;B9735;386
Entity A;C1299;322
Entity A;D1918;
Entity A;E0702;
Entity A;F1290;
Entity B;A1136;1982
Entity B;B0765;988
Entity B;C8099;522
Entity B;D3938;
Entity B;E5722;
Entity B;F3596;
[download]

findvalue('./quantity/text()')//"0";
[download]

You can use Data::Dumper to see what your variables contain. For articles without quantity, $amount is the empty string q(), which means it is defined. If you replace // with ||, you will get the desired behaviour.

Thank you choroba it works this way! Again I've learned on PM!
... now trying this "in real life" with the complete file.
Thanks again.
VE