How to insert HTML to PHP DOMNode?

You can use

  • DOMDocumentFragment::appendXML — Append raw XML data

Example:

// just some setup
$dom = new DOMDocument;
$dom->loadXml('<html><body/></html>');
$body = $dom->documentElement->firstChild;

// this is the part you are looking for    
$template = $dom->createDocumentFragment();
$template->appendXML('<h1>This is <em>my</em> template</h1>');
$body->appendChild($template);

// output
echo $dom->saveXml();

Output:

<?xml version="1.0"?>
<html><body><h1>This is <em>my</em> template</h1></body></html>

If you want to import from another DOMDocument, replace the three lines with

$tpl = new DOMDocument;
$tpl->loadXml('<h1>This is <em>my</em> template</h1>');
$body->appendChild($dom->importNode($tpl->documentElement, TRUE));

Using TRUE as the second argument to importNode will do a recursive import of the node tree.


If you need to import (malformed) HTML, change loadXml to loadHTML. This will trigger the HTML parser of libxml (what ext/DOM uses internally):

libxml_use_internal_errors(true);
$tpl = new DOMDocument;
$tpl->loadHtml('<h1>This is <em>malformed</em> template</h2>');
$body->appendChild($dom->importNode($tpl->documentElement, TRUE));
libxml_use_internal_errors(false);

Note that libxml will try to correct the markup, e.g. it will change the wrong closing </h2> to </h1>.


It works with another DOMDocument for parsing the HTML code. But you need to import the nodes into the main document before you can use them in it:

$newDiv = $dom->createElement('div');
$tmpDoc = new DOMDocument();
$tmpDoc->loadHTML($str);
foreach ($tmpDoc->getElementsByTagName('body')->item(0)->childNodes as $node) {
    $node = $dom->importNode($node, true);
    $newDiv->appendChild($node);
}

And as a handy function:

function appendHTML(DOMNode $parent, $source) {
    $tmpDoc = new DOMDocument();
    $tmpDoc->loadHTML($source);
    foreach ($tmpDoc->getElementsByTagName('body')->item(0)->childNodes as $node) {
        $node = $parent->ownerDocument->importNode($node, true);
        $parent->appendChild($node);
    }
}

Then you can simply do this:

$elem = $dom->createElement('div');
appendHTML($elem, '<h1>Hello world</h1>');

As I do not want to struggle with XML, because it throws errors faster and I am not a fan of prefixing an @ to prevent error output. The loadHTML does the better job in my opinion and it is quite simple as that:

$doc = new DOMDocument();
$div = $doc->createElement('div');

// use a helper to load the HTML into a string
$helper = new DOMDocument();
$helper->loadHTML('<a href="#">This is my HTML Link.</a>');

// now the magic!
// import the document node of the $helper object deeply (true)
// into the $div and append as child.
$div->appendChild($doc->importNode($helper->documentElement, true));

// add the div to the $doc
$doc->appendChild($div);

// final output
echo $doc->saveHTML();

Here is simple example by using DOMDocumentFragment:

$doc = new DOMDocument();
$doc->loadXML("<root/>");
$f = $doc->createDocumentFragment();
$f->appendXML("<foo>text</foo><bar>text2</bar>");
$doc->documentElement->appendChild($f);
echo $doc->saveXML();

Here is helper function for replacing DOMNode:

/** 
 * Helper function for replacing $node (DOMNode) 
 * with an XML code (string) 
 * 
 * @var DOMNode $node 
 * @var string $xml 
 */ 
public function replaceNodeXML(&$node, $xml) { 
  $f = $this->dom->createDocumentFragment(); 
  $f->appendXML($xml); 
  $node->parentNode->replaceChild($f,$node); 
}

Source: Some old "PHP5 Dom Based Template" article.

And here is another suggestion posted by Pian0_M4n to use value attribute as workaround:

$dom = new DomDocument;

// main object
$object = $dom->createElement('div');

// html attribute
$attr = $dom->createAttribute('value');
// ugly html string
$attr->value = "<div>&nbsp; this is a really html string &copy;</div><i></i> with all the &copy; that XML hates!";
$object->appendChild($attr);

// jquery fix (or javascript as well)
$('div').html($(this).attr('value')); // and it works! 
$('div').removeAttr('value'); // to clean-up

No ideal, but at least it works.