Question

What are reasons why an XPath query could only return the last match? I am running a query on an HTML fragment which clearly has multiple <a name="..."> tags, but the XPath query will only return one element, which happens to be the last element.

function extract($html) {
    // This test shows that the retrieved HTML fragment indeed contains multiple anchor tags
    echo "<textarea>".$html."</textarea>";

    // parse the data
    $dom = new DomDocument();
    @$dom->loadHTML($html); // we use @$dom to suppress some warnings
    $xpath = new DOMXPath($dom);

    // find the html code for the post
    $query = "//a[contains(@name, 'post')]"; 
    $rows = $xpath->query($query);

    // This will return 1
    echo "Elements found: " . count($rows);

    ...
}

Answer 1

You need to probe $rows->length. The reason is that $rows is an instance of DOMNodeList (an object that holds a list of DOMNode instances). And since DOMNodeList doesn’t implement the interface Countable, it cannot be probed by count() the way you’d expect. It simply returns 1, because it is a single object, not the amount of DOMNodes it has aggregated.

And so, the result of your query does not return only the last match. It returns them all, and you can iterate through them with foreach, like so:

foreach( $rows as $row )
{
    // do something with $row (instance of DOMNode)
}

… or with a for loop, like so:

for( $i = 0, $len = $rows->length; $i < $len; $i++ )
{
    $row = $rows->item( $i );
    // do something with $row (instance of DOMNode)
}

… etc.