PHP DOMElement::getElementsByTagName - Anyway to get just the immediate matching children?

Question

is there a way to retrieve only the immediate children found by a call to DOMElement::getElementsByTagName? For example, I have an XML document that has a category element. That category element has sub category elements (which have the same structure), like:

<category>
    <id>1</id>
    <name>Top Level Category Name</name>
    <subCategory>
        <id>2</id>
        <name>Sub Category Name</name>
    </subCategory>
    ...
</category>

If I have a DOMElement representing the top level category,

$topLevelCategoryElement->getElementsByTagName('id');

will return a list with the nodes for all 'id' elements, where I want just the one from the top level. Any way to do this outside of using XPath?

Answer 1

I'm afraid not. You'll have to iterate through the children or use XPath.

for ($n = $parent->firstChild; $n !== null; $n = $n->nextSibling) {
    if ($n instanceof DOMElement && $n->tagName == "xxx") {
        //...
    }
}

Example with XPath and your XML file:

$xml = ...;
$d = new DOMDocument();
$d->loadXML($xml);
$cat = $d->getElementsByTagName("subCategory")->item(0);
$xp = new DOMXpath($d);
$q = $xp->query("id", $cat); //note the second argument
echo $q->item(0)->textContent;

gives 2.

Answer 2

Something like this should do

/**
 * Traverse an elements children and collect those nodes that
 * have the tagname specified in $tagName. Non-recursive
 *
 * @param DOMElement $element
 * @param string $tagName
 * @return array
 */
function getImmediateChildrenByTagName(DOMElement $element, $tagName)
{
    $result = array();
    foreach($element->childNodes as $child)
    {
        if($child instanceof DOMElement && $child->tagName == $tagName)
        {
            $result[] = $child;
        }
    }
    return $result;
}

edit: added instanceof DOMElement check

Answer 3

I don't use PHP, but if PHP actually implements the DOM API as specified the W3C, there is required to be a childNodes property on any Node object. You should be able to iterate over all of the direct children and test their tag name to see if they're the one you're looking for. Depending on what your tree looks like, this may be slower than getting all the elements by tag name and testing their tree position, or it may be significantly faster.