PHP DOM Trouble - Selectively Deleting Nodes

[mmcdermo]

********Edit**********

If you're looking for a working function that selectively deletes nodes based on
a) the text of a child node, go here
b) the attribute of a node, go here

********/Edit**********

My goal here is to create a function that deletes nodes from an XML file if it meets certain criterion passed to the function.

Here's the basic structure of the XML file I'm trying to delete nodes from:
<?xml version="1.0" encoding="ISO-8859-1"?>
<tier1_obj>
<tier2_obj>
<tier3_obj>Some Text</tier3_obj>
<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
<tier2_obj>
<tier3_obj>Other Text</tier3_obj>
<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
<tier2_obj>
<tier3_obj>Some Text</tier3_obj>
<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
</tier1_obj>

So, let's say that we want to delete any tier2_obj's that have a tier3_obj thats textContent property is "Some Text". Here's the function I came up with to do this [Note the $i_position variable. I use this because once you delete a node from $node_list, the other nodes' indicies are reduced by 1. Every time I delete a node, I subtract one from $i_position] :

    	  function selectivelyDeleteNodes ($tagname, $cNode, $textContentValue) {
                 // Find all elements with specefied $tagname
                 $xml_file = new DOMDocument;
                 $xml_file->load($some_xml_file);
                 $xml_file_delm = $xml_file->documentElement;
                 $node_list= $xml_file_delm->getElementsByTagName("$tagname");
                 $num_nodes = $node_list->length; //Number of matching nodes

             for ($i = 0; $i < $num_nodes; $i++ ) {
                  if(!isset($i_position)){$i_position = 0;} //For first run, initiate var i_position with value 0

                    echo "<hr/><br/>Run Through $i:$i_position<br/>";
                    $parent_node = $node_list->item($i_position)->parentNode;
                    echo "Parent Node = &lt;"; echo $parent_node->nodeName; echo "&gt;<br/>";
                    $child_to_delete = $node_list->item($i_position);
                    echo "Child Node = &lt;"; echo $child_to_delete->nodeName; echo "&gt;<br/>";
                    if($child_to_delete->$cNode->textContent == $textContentValue){
                      $parent_node->removeChild($child_to_delete);
                      $i_position--;
                    }
                    $i_position++;
            }
}

Whenever I run this function, it never deletes the <tier2_obj> even when I modify the if statement to something like if(1=1).. So I believe my problem comes down to the line:

$parent_node->removeChild($child_to_delete);

Although it's really quite hard for me to tell.

I'd appreciate any help on the subject, if even that help is just a link to somewhere that I might get the inspiration to figure out what's wrong with my code here 🙂.
Thank you for your valuable time,
~Morgan McDermott.

EDIT: Apologies to those who read this thread before 6:17PM PST. I forgot to change some of the variable names from when the function was part of a larger class ^_.

[sneakyimp]

What does your output look like?

[mmcdermo]

From the echo commands I built into the function, my output is this:

Run Through 0:0
Parent Node = <tier2_obj>
Child Node = <tier3_obj>

---

Run Through 1:0
Parent Node = <tier2_obj>
Child Node = <tier3_obj>

What I get from this output is that the script is selecting the proper child nodes, and is running through the proper number [$node_list->length] of times. However, when I look at the XML output in a file [ In testing, I added $xml_file->save($file);], the nodes are still very much there.

😕 I'm very much confused by this... Thank you for reminding me to post my output, though!

[sneakyimp]

can you run this? It will give us a little more info:

function selectivelyDeleteNodes ($tagname, $cNode, $textContentValue) {
	// Find all elements with specefied $tagname
	$xml_file = new DOMDocument;
	$xml_file->load($some_xml_file);
	$xml_file_delm = $xml_file->documentElement;
	$node_list= $xml_file_delm->getElementsByTagName("$tagname");
	$num_nodes = $node_list->length; //Number of matching nodes

for ($i = 0; $i < $num_nodes; $i++ ) {
	if(!isset($i_position)){$i_position = 0;} //For first run, initiate var i_position with value 0

	echo "<hr/><br/>Run Through $i:$i_position<br/>";
	$parent_node = $node_list->item($i_position)->parentNode;
	echo "Parent Node = &lt;"; echo $parent_node->nodeName; echo "&gt;<br/>";
	$child_to_delete = $node_list->item($i_position);
	echo "Child Node = &lt;"; echo $child_to_delete->nodeName; echo "&gt;<br/>";
echo 'textContent of $child_to_delete:' . $child_to_delete->$cNode->textContent . '<br>';
echo '$textContentValue:' . $textContentValue . '<br>';
echo 'comparison result:' . ($child_to_delete->$cNode->textContent == $textContentValue) . '<br>';
		if($child_to_delete->$cNode->textContent == $textContentValue){
echo ('DELETING<br>');
			$parent_node->removeChild($child_to_delete);
			$i_position--;
		}
		$i_position++;
	} // for i...

print_r($node_list);


} // selectivelyDeleteNodes()

[mmcdermo]

Alright, apologies in advance, but this post is going to be rather long-winded.

First of all, in order to get the function you gave me to run properly, I had to remove '$cNode' from everything, as it didn't want to function properly. I believe that this was part of my problem... I'm not sure if the PHP DOM objects allow you to access eachother like
$node1->node2->textContent

Or at least, that seems to be the problem here. When I left in '$cNode->', the textContent property of the object [ $child_to_delete->$cNode->textContent ] would always be blank.

The solution to get your script to run, then, was to remove $cNode and change the functionality slightly. I changed it to delete any node whose textContent matched the third function argument [ $textContentValue ].

Modified function:

    function selectivelyDeleteNodes ($tagname, $cNode, $textContentValue) {
    // Find all elements with specefied $tagname
    $xml_file = new DOMDocument;
    $xml_file->load($some_xml_file);
    $xml_file_delm = $xml_file->documentElement;
    $node_list= $xml_file_delm->getElementsByTagName("$tagname");
    $num_nodes = $node_list->length; //Number of matching nodes

for ($i = 0; $i < $num_nodes; $i++ ) {
    if(!isset($i_position)){$i_position = 0;} //For first run, initiate var i_position with value 0

    echo "<hr/><br/>Run Through $i:$i_position<br/>";
    $parent_node = $node_list->item($i_position)->parentNode;
    echo "Parent Node = &lt;"; echo $parent_node->nodeName; echo "&gt;<br/>";
    $child_to_delete = $node_list->item($i_position);
    echo "Child Node = &lt;"; echo $child_to_delete->nodeName; echo "&gt;<br/>";
echo 'textContent of $child_to_delete:' . $child_to_delete->textContent . '<br>';  //$cNode->
echo '$textContentValue:' . $textContentValue . '<br>';
echo 'comparison result:' . ($child_to_delete->textContent == $textContentValue) . '<br>'; //  $cNode->
        if($child_to_delete->textContent == $textContentValue){  //->$cNode
echo ('DELETING<br>');
            $parent_node->removeChild($child_to_delete);
            $i_position--;
       }
        $i_position++;
    } // for i...

print_r($node_list);
$file = "$this->source"."1";
 $xml_file->save($file);

}

The output from this function when run with arguments "tier3_obj", '', 'Some Text'
is :
---------------Output-------------
Run Through 0:0
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete:Some Text
$textContentValue:Some Text
comparison result:1
DELETING

---

Run Through 1:0
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete😮ther Text
$textContentValue:Some Text
comparison result:

---

Run Through 2:1
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete😮ther Text
$textContentValue:Some Text
comparison result:

---

Run Through 3:2
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete😮ther Text
$textContentValue:Some Text
comparison result:

---

Run Through 4:3
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete:Some Text
$textContentValue:Some Text
comparison result:1
DELETING

---

Run Through 5:3
Parent Node = <tier2_obj>
Child Node = <tier3_obj>
textContent of $child_to_delete😮ther Text
$textContentValue:Some Text
comparison result:
DOMNodeList Object ( )

--------------------/Output-----------------------

In the file, the contents were as follows:
<?xml version="1.0" encoding="ISO-8859-1"?>
<tier1_obj>
<tier2_obj>

<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
<tier2_obj>
<tier3_obj>Other Text</tier3_obj>
<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
<tier2_obj>

<tier3_obj>Other Text</tier3_obj>
</tier2_obj>
</tier1_obj>

So, removing the extra node in the (reference? What do you call this 😃 ?) made it functional... Now, I'm wondering if I could have the function remove a node if a certain attribute [like an ID attribute] were equal to one of the function's arguments. I'm also very, very curious as to why accessing a node property like $node1->node2->property doesn't work.

Thanks a lot for your help sneakyimp. I have a fair idea of what my problem was, although all this PHP DOM stuff is still swirling around in my head as I try to get a firm grasp on it. Anyways, if anyone would be so kind as to help me figure out why I couldn't access node properties like that, and if accessing attributes WOULD work instead, I'd be greatly appreciative.

:evilgrin: Thanks again, sneaky. I feel as if I have a firmer grip on what I must do, now.

[mmcdermo]

I've decided that the easiest way to do this would be to delete a node if a certain attribute (like an id) matched an argument passed to the function. Although you could create a function that deletes a node if a child node of that node has a certain textContent property by looping through all nodes, finding a node with that textContent and then deleting its parent, deleting by attribute is simpler to code.

Here's the function I've come up with:

function selectivelyDeleteNodes ($tagname, $attrName, $attrValue) {
    // Find all elements with specefied $tagname
    $xml_file = new DOMDocument;
    $xml_file->load($some_xml_file);
    $xml_file_delm = $xml_file->documentElement;
    $node_list= $xml_file_delm->getElementsByTagName("$tagname");
    $num_nodes = $node_list->length; //Number of matching nodes

for ($i = 0; $i < $num_nodes; $i++ ) {
    if(!isset($i_position)){$i_position = 0;} //For first run, initiate var i_position with value 0

    $parent_node = $node_list->item($i_position)->parentNode;
    $child_to_delete = $node_list->item($i_position);

    if($child_to_delete->getAttribute("$attrName") == $attrValue){ 
        $parent_node->removeChild($child_to_delete);
        $i_position--;
   }
    $i_position++;
} // for i...

$file = "somefile.extension";
 $xml_file->save($file);


} // selectivelyDeleteNodes()

[mmcdermo]

Alright, I made a function this morning to solve my ORIGINAL problem, although I will probably use the first solution of deleting by attribute instead.

Thanks again to the PHPBuilder community and sneakyimp... I just hope that if anyone has this problem in the future, they find their solution in this thread. This function is useful if you have an XML layout something like

<element1>
<id>100</id>
<information>....</information>
</element1>
<element2>
<id>101</id>
...
...
...

To recap, this function deletes a node using removeChild if it has a child node thats textContent property is equal to $textContentValue, which is the second argument of the function.

function deleteNodeByChild ($childtagname, $textContentValue, $XML_string) {
              // Find all elements with specefied $tagname
              $xml_file = new DOMDocument;
              $xml_file->loadXML($XML_string);
              $xml_file_delm = $xml_file->documentElement;
              $node_list= $xml_file_delm->getElementsByTagName("$childtagname");
              $num_nodes = $node_list->length; //Number of matching nodes

          for ($i = 0; $i < $num_nodes; $i++ ) {
              if(!isset($i_position)){$i_position = 0;} //For first run, initiate var i_position with value 0

              $child_to_delete = $node_list->item($i_position);
              $parent_node = $node_list->item($i_position)->parentNode;
              $parent_parent_node = $parent_node->parentNode;

              if($child_to_delete->textContent == $textContentValue){
                  $parent_parent_node->removeChild($parent_node);
                  $i_position--;
             }

              $i_position++;
          } // for i...
          $xml_file->formatOutput = true;
          return $xml_file->saveXML();
  } // deleteNodeByChild()

Note that the solution here was to search for and get the most deeply nested child node [ id ] instead of the node we actually wanted to delete [say, element1] and then work back UP to the PARENT of the node we wanted to delete, and then call removeChild.