HTML::Tree: get value of an element

[Ratazong]

Dear Monks!

I'm playing with html-parsing (using HTML::Tree), and am stuck with parsing the following snippet:

<span class="name color1">ImportantText1</span>
<span class="name color2"><span class="level">1000</span>ImportantText
+2</span>

I try to parse this with the code below (just a part of the script to show you the idea), and finally want to have the ImportantText1 or ImportantText2 in the variable $name. However I end with the value 1000ImportantText2 in the second case. And somehow I cannot find the part of the documentation which shows me how to replace the as_trimmed_text() so the values of the sub-elements are ignored.

Please enlighten me!

Rata

my @subsub = $body->look_down (_tag => "span");
foreach my $sss (@subsub) {
    my $c     = $sss->attr("class");
    if ($c =~ /name color/)   { $name  = $sss->as_trimmed_text(); }
}

[poj]

Try

#!perl
use strict;
use HTML::TreeBuilder;

my $body = HTML::TreeBuilder->new_from_file(\*DATA);
my @subsub = $body->look_down (_tag => "span");

foreach my $sss (@subsub) {
   my $c = $sss->attr("class");
   if ($c =~ /name color/){ 

     # ignore sub elements
     for ( $sss->descendants() ){
       $_->detach();
     };
     
     my $name = $sss->as_trimmed_text();
     print $name."\n";
   }
}

__DATA__
<span class="name color1">ImportantText1</span>
<span class="name color2"><span class="level">1000</span>ImportantText
+2</span>

poj

[Ratazong]

Thanks for your suggestion! I tested it, and it works fine. :-)

Nevertheless, I solved the problem another way (as I don't like the idea of destroying my tree): I now additionally create the texts for the sub-elements, and then remove them from the top-element.

This solution also doesn't make me happy (it is somehow waste) - however I don't have any problems with runtime or memory-usage now, so it doesn't hurt. And the additional amount of electrons moving around heats up the air in the room - which is not to bad in winter ;-)

Rata