It’s odd. I don’t think about think about XPath from one month to the next. But once in a while, when my usual solutions have all come up blank. Ta-Da XPath to the rescue!

Recently as part of a site I was working on, the design basically required that I inject a block of content into a nested list (part of an elaborate menu actually) – bit of a fiddle because modifying the code that generated the list was not an option.

My first avenues of attack were just simple str_replace() and a regex replace, but I just couldn’t get it work consistently – there were two many variables – additional attributes, white-space etc. The two constants were that the code fragment validated as xhtml (hence xml) and I would always have a class that I could use as a hook.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

 
// $menu - fragment of html consisting of nested UL's 
 
$xml = simplexml_load_string($menu);
 
$nodes = $xml->xpath('//*[@class = "current"]');
 
if(!empty($nodes)){
 
	$nodes[0]->addChild('div', 'text_to_replace');
	$menu = $xml->asXML();
	echo str_replace('text_to_replace', $str, $menu);
 
} else {
	echo $menu;
}

//*[@class = "current"] finds nodes at any level with where the class attribute contains ‘current’.

node[0] is the first instance of a node with this attribute and inject some place-holder text.

$menu = $xml->asXML(); our modified list.

Hey presto!