Diese Seite bei php.net
The PHP manual text and comments are covered by the Creative Commons Attribution 3.0 License © the PHP Documentation Group - Impressum - mail("TO:Reinhard Neidl",...)
(PHP 5)
DOMNode::removeChild — Removes child from list of children
This functions removes a child from a list of children.
The removed child.
If the child could be removed the functions returns the old child.
Raised if this node is readonly.
Raised if oldnode is not a child of this node.
The following example will delete the chapter element of our XML document.
Beispiel #1 Removing a child
<?php
$doc = new DOMDocument;
$doc->load('book.xml');
$book = $doc->documentElement;
// we retrieve the chapter and remove it from the book
$chapter = $book->getElementsByTagName('chapter')->item(0);
$oldchapter = $book->removeChild($chapter);
echo $doc->saveXML();
?>
Das oben gezeigte Beispiel erzeugt folgende Ausgabe:
<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN"
"http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd">
<book id="listing">
<title>My lists</title>
</book>
Beispiel #2 Preserving the parent node's namespace URI
<?php
$doc = new DOMDocument;
$doc->load('book.xml');
$book = $doc->documentElement;
// we retrieve the chapter and remove it from the book
$chapter = $book->getElementsByTagName('chapter')->item(0);
// copy the namespace URI
$nsuri = $book->namespaceURI;
// Delete the child node
$book->removeChild($chapter);
// paste the namespace URI back into the parent node
$book->namespaceURI = $nsuri;
?>
Hinweis:
After calling this method, the DOMNode::$namespaceURI property on the parent node will be reset to NULL. Above example shows how to work around it.
For the loop bug, you have to use the parent node of the node itself :
<?php
$nodeList = $doc->getElementsByTagName('p');
for ($i = 0; $i < $nodeList->length; $i++) {
try {
$old_child = $nodeList->item($i)->parentNode->removeChild($nodeList->item($i));
} catch (DOMException $e) {
echo $e->getMessage();
}
}
?>
If you try to remove children of a nested node (e.g. a <ul> list that contains another <ul>) and only remove the first level of child nodes, you'll get a messed-up node object.
Upon re-use of the messed-up node you'll get an error message that says: "Fatal error: Call to a member function blahblah() on a non-object".
To remove ALL child nodes, use this recursive function instead:
<?php
/**
* Helper function that removes all child nodes of $node recursively
* (i.e. including child nodes of the child nodes and so forth).
*/
function remove_children(&$node) {
while ($node->firstChild) {
while ($node->firstChild->firstChild) {
remove_children($node->firstChild);
}
$node->removeChild($node->firstChild);
}
}
?>
Just a note, if you remove a node the whitespace it occupied will remain if you save the file.
To avoid this, use $document->preserveWhiteSpace = false;
For Example:
<?php
$document = new DomDocument();
$document->preserveWhiteSpace = false;
?>
You can't remove DOMNodes from a DOMNodeList as you're iterating over them in a foreach loop. For example...
<?php
$domNodeList = $domDocument->getElementsByTagname('p');
foreach ( $domNodeList as $domElement ) {
// ...do stuff with $domElement...
$domElement->parentNode->removeChild($domElement);
}
?>
... will seemingly leave the internal iterator on the foreach out of wack and results will be quite strange. Though, making a queue of items to remove seems to work. For example...
<?php
$domNodeList = $domDocument->getElementsByTagname('p');
$domElemsToRemove = array();
foreach ( $domNodeList as $domElement ) {
// ...do stuff with $domElement...
$domElemsToRemove[] = $domElement;
}
foreach( $domElemsToRemove as $domElement ){
$domElement->parentNode->removeChild($domElement);
}
?>
These two functions might be helpful to anyone trying to delete a node and all of its children:
<?php
function deleteNode($node) {
deleteChildren($node);
$parent = $node->parentNode;
$oldnode = $parent->removeChild($node);
}
function deleteChildren($node) {
while (isset($node->firstChild)) {
deleteChildren($node->firstChild);
$node->removeChild($node->firstChild);
}
}
?>
Or something like:
<?php
while ($node->childNodes->length)
$node->removeChild($node->firstChild);
?>
Note that iterating through the childNodes array and removing those children will stop the iteration.
For example:
foreach($node->childNodes as $child)
{
if(iWantToDeleteThisNode($child))
$child->parentNode->removeChild($child);
}
Will not work. Note that in removing the node, the childNodes array gets rebuilt (it seems) and so only the first item will be removed. In order to properly remove all the children you want to remove, you will need to do something like the following:
$nodesToDelete = array();
foreach($node->childNodes as $child)
{
if(iWantToDeleteThisNode($child))
$nodesToDelete[] = $child;
}
foreach($nodesToDelete as $node)
{
$node->parentNode->removeChild($node);
}
I believe this is actually more efficient than the first snippet would be (if it worked) but obviously I cannot benchmark it.
I used the following to delete an entire node using the child element:
<?php
$childElement->parentNode->parentNode-> removeChild($childElement->parentNode);
?>
I could not have thought of it without the help of everyone who contributed to this thread.
Thanks!
(I had to introduce the extra space between parentNode->removeChild to get rid of the wordwrap() warning while posting this note)
At the time of writing, I suppose rightfully, removeChild() removes only the selected node, but when you remove an element, it's child elements are not removed. If you want to achieve that, replaceChild() is the solution.
The following should remove all descendants of the $node DOMNode, regardless of it's name:
<?php
$node->replaceChild(new DOMElement($node->nodeName), $node);
?>
If you're replacing the root element, you must explicitly state that with $node->documentElement as the second argument.
The previous example is not correct; it only works when the tags are direct childs of the top-level node.
Better is:
while($elem = $dom->getElementsByTagName("name")->item(0)) {
$elem->parentNode->removeChild($elem);
}
remove all children of a given node (simple way):
$node->nodeValue = '';
Back again on removing childs and iterators robustness.
Things get a bit more complicated when you only want to remove 'some' nodes according to a certain condition. Then you can't just remove the first one repeatedly.
The trick is to copy the content of the node list into a more robust collection than DOMNodeList, I name array!
The following piece of code will, for instance, remove all empty child nodes:
<?php
// Copy childNodes array
$childNodes = array();
foreach($node->childNodes as $childNode) {
$childNodes[] = $childNode;
}
// Browse with the copy
foreach ($childNodes as $childNode) {
if (!$childNode->hasChildNodes()); {
$childNode->parentNode->removeChild($childNode);
}
}
?>
can we directly remove a node which has some children or do we need to remove the children first.
while($mytn->hasChildNodes())
{
$mytn->removeChild($mytn->childNodes->item(0));
}
echo $j;//this shows correct no of children
echo $mytn->childNodes->length;//after removing it
// shows 0..correct
$dom->save(net.xml);
but the changes r not reflected in the xml file
This might be obvious, but might help someone as well...
If you simply have a node that you need to remove (perhaps from an xpath) and don't know the parent node offhand (the examples seem to assume you have the parent in a variable), then you can do something easy like this...
$node->parentNode->removeChild($node);
parentNode is a scalar property of the Element.
Hope that helps.
--
Blaine Garrett
Executive Director of the Art Attack
http://theartattack.org
You may also use the hasChild function:
<?php
while($node->hasChildNodes()) {
$node->removeChild($node->childNodes->item(0));
}
?>
When you remove a childNode, the next node becomes the first one!
For those who don't understand >sbeam at onsetcorps dot net on 02-Feb-2005 12:07< 'hack', here's a little discussion:
First but *wrong* try would be:
<?php
foreach ($parent->childNodes as $child) {
$parent->removeChild($child);
?>
This doesn't work because DOM tree id modified on-the-fly and this confuses foreach loop.
The idea behind sbeam's trick is that after removing the first item in the first iteration, the second item in childNodes nodelist immediately becomes the first item. That's why we must _always_ remove the first child. Here's another implementation:
<?php
$count = $parent->childNodes->length;
for ($i = 0; $i < $count; $i++) {
$oldNode = $parent->removeChild($parent->childNodes->item(0)); // !!! not item($i) !!!
}
?>
remove all children of a given node:
while ($parent->childNodes->length) {
$parent->removeChild($parent->childNodes->item(0));
}
Diese Seite bei php.net