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)
DOMXPath::query — Evaluates the given XPath expression
Executes the given XPath expression.
The XPath expression to execute.
The optional contextnode can be specified for doing relative XPath queries. By default, the queries are relative to the root element.
The optional registerNodeNS can be specified to disable automatic registration of the context node.
Returns a DOMNodeList containing all nodes matching the given XPath expression. Any expression which does not return nodes will return an empty DOMNodeList.
If the expression is malformed or the contextnode is invalid, DOMXPath::query() returns FALSE.
| Version | Beschreibung |
|---|---|
| 5.3.3 | The registerNodeNS parameter was added. |
Beispiel #1 Getting all the english books
<?php
$doc = new DOMDocument;
// We don't want to bother with white spaces
$doc->preserveWhiteSpace = false;
$doc->Load('book.xml');
$xpath = new DOMXPath($doc);
// We starts from the root element
$query = '//book/chapter/para/informaltable/tgroup/tbody/row/entry[. = "en"]';
$entries = $xpath->query($query);
foreach ($entries as $entry) {
echo "Found {$entry->previousSibling->previousSibling->nodeValue}," .
" by {$entry->previousSibling->nodeValue}\n";
}
?>
Das oben gezeigte Beispiel erzeugt folgende Ausgabe:
Found The Grapes of Wrath, by John Steinbeck Found The Pearl, by John Steinbeck
We can also use the contextnode parameter to shorten our expression:
<?php
$doc = new DOMDocument;
$doc->preserveWhiteSpace = false;
$doc->load('book.xml');
$xpath = new DOMXPath($doc);
$tbody = $doc->getElementsByTagName('tbody')->item(0);
// our query is relative to the tbody node
$query = 'row/entry[. = "en"]';
$entries = $xpath->query($query, $tbody);
foreach ($entries as $entry) {
echo "Found {$entry->previousSibling->previousSibling->nodeValue}," .
" by {$entry->previousSibling->nodeValue}\n";
}
?>
If the query() function seems to ignore your $contextnode, and instead returns all the tags in the document, try to use a relative path (use a . in front of the query):
<?php
$xml = "<?xml version='1.0' encoding='UTF-8'?>
<test>
<tag1>
<uselesstag>
<tag2>test</tag2>
</uselesstag>
</tag1>
<tag2>test2</tag2>
</test>";
$dom = new DomDocument();
$dom->loadXML($xml);
$xpath = new DomXPath($dom);
$tag1 = $dom->getElementsByTagName("tag1")->item(0);
echo $xpath->query("//tag2")->length; //output 2 -> correct
echo $xpath->query("//tag2", $tag1)->length; //output 2 -> wrong, the query is not relative
echo $xpath->query(".//tag2", $tag1)->length; //output 1 -> correct (note the dot in front of //)
?>
See that i couldn't use $xpath->query("tag2", $tag1) as per the documentation, since "tag2" is not a direct child of "tag1".
I don't know why this note was deleted, i just tested it and it's correct.
It's not a bug, it's simply not written in the documentation.
I found this useful for building page templates
<?php
$xsl = new DOMDocument;
$xsl->load('layout.xsl');
// Set the <xsl:include> href attribute, the inner stylesheet to include in this layout
$xpath = new DomXPath($xsl);
$res = $xpath->query('//xsl:include');
$res->item(0)->setAttribute('href','page.xsl');
$xsl->save('media/xsl/layout.xsl');
?>
For XPath escaping use the following method (of course it could be more efficient).
<?php
public function xpathescape($string)
{
$result = 'concat(';
for($i=0, $j=strlen($string); $i<$j; ++$i)
{
if($i > 0)
$result .= ",";
if($string[$i] == '\'')
$result .= "\"".$string[$i]."\"";
else
$result .= '\''.$string[$i].'\'';
}
$result .= ')';
return $result;
}
?>
Use it this way:
<php
$xpath->query('//example[sub='.xpathescape($acomplexstring).']');
?>
If you're wondering, like I was, why your XPath queries are not returning any of the new DOMElements you create in your (X)HTML documents, and only the ones originally loaded in with (for example) loadXML(), this is why; if you're doing things right, you have registered the nameSpace 'html' after creating your DOMXPath object thus:
<?php
class XPathQueryLength {
private $nameSpace = '';
function __construct(DOMDocument $doc) {
$this->xpath = new DOMXPath($this->doc);
$this->xpath->registerNamespace(
'html','http://www.w3.org/1999/xhtml' );
}
function queryLength($query) {
return $this->xpath->query($query)->length;
}
}
?>
...but don't forget that when adding new elements to the above DOMDocument $doc, to use createElementNS() instead of createElement(), otherwise you'll have this problem:
<?php
//$doc is a previously loaded XHTML document containing a normal html, head and body structure
//$body is the first selected tag using $doc->getElementsByTagName('body');
$pTag = $doc->createElement('p','This is a new paragraph!');
$body->appendChild($pTag);
$xPath = new XPathQueryLength($doc);
print $xPath->queryLength('//html:p');
output: 0
print $xPath->queryLength('//p');
output: 1
?>
So do this instead:
<?php
//$doc is a previously loaded XHTML document containing a normal html, head and body structure
//$body is the first selected tag using $doc->getElementsByTagName('body');
$pTag = $doc->createElementNS('http://www.w3.org/1999/xhtml','p','This is a new paragraph!');
$body->appendChild($pTag);
$xPath = new XPathQueryLength($doc);
print $xPath->queryLength('//html:p');
output: 2
print $xPath->queryLength('//p');
output: 0
?>
The resulting XHTML file from both example scripts looks much like this:
<html>
<head></head>
<body>
<p>This is a hardcoded paragraph.</p>
<p>This is a new paragraph!</p>
</body>
</html>
...so you would think that a paragraph is a paragraph is a paragraph, since you never see the prefix, as in "<html:p>This is a new paragraph!</html:p>".
This may seem glaringly obvious, but I was writing a class that converts CSS queries into XPath queries, and the fact that a namespace had been registered was rather buried in the code.
We love the DOM, the DOM is good to us.
I've searched the entire web looking for a way to update / modify/ change/ alter the elements of an xml file and found NOTHING!
So here it is, the defninitive way to "Change XML elements with PHP" rather than adding / appending new ones. This uses XPATH:
<?php
// Create a DOMDocument instance
$xml = new DOMDocument;
// Ignore whitespace between nodes (default: true)
$xml->preserveWhiteSpace = false;
$file='about.xml';
// Load the XML data source
$xml->Load($file);
$xpath = new DOMXPath($xml);
$query='/regions/branch';
$entries = $xpath->query($query);
foreach ($entries as $entry)
{
$entry->firstChild->nodeValue="like this!";
echo $entry->firstChild->nodeValue;
}
$xml->save($file);
?>
Unfortunately PHP's DOM extension doesn't support use of:
<?xml-stylesheet type="text/xsl" ... ?>
processing instruction.
Here is an example, how to implement it using XPath query and extending DOMDocument by a method output().
<?php
// This simple function adds missing direct usage of anonymous instances
// in PHP5's reference model
function a($var) {
return $var;
}
// Extended DOMDocument class
class MyDOMDocument extends DOMDocument
{
public function output()
{
$stylesheets = array();
$PIs = a(new DOMXPath($this))
->query('/processing-instruction("xml-stylesheet")');
foreach($PIs as $PI)
{
// This might be implemented cleaner by regular parsing
// of DOMProcessingInstruction::data property
if(ereg('type *= *"text/xsl" +href *= *"([^"]+)"', $PI->data, $mem))
{
// Here should be verified, that XSL file exists.
a($stylesheets[] = new DOMDocument())->load($mem[1]);
}
}
if($stylesheets)
{
$processor = new XSLTProcessor();
foreach($stylesheets as $stylesheet)
$processor->importStylesheet($stylesheet);
return $processor->transformToDoc($this);
}
// If no stylesheet instructions present, return self directly
else return $this;
}
}
?>
Usage:
<?php
$document = new MyDOMDocument();
$document->load('my.xml');
echo $document->output()->saveXML();
?>
With following file my.xml:
<?xml version="1.0" ?>
<?xml-stylesheet type="text/xsl" href="my.xsl" ?>
<my-root />
and existing file my.xsl that code will transform the xml file using my.xsl and output the result.
The order of nodes in the return value is not guaranteed.
When my code was on an old server, the returned DOMNodeList was in document order. On the new server, the returned DOMNodeList is in a consistent order, but it is not in document order.
PHP passes this function call off to the xmlXPathEvalExpression() function in libxml. That function in libxml only accepts two arguments - the same two this PHP function accepts. There must have been a change in the libxml version from the old server to the new server, and that libxml behaves differently.
This would be okay if PHP had a way to compare nodes so I can resort the nodes manually, but there is not.
So, there is no guaranteed way to get an ordered list of nodes like DOM 3 XPath provides.
Note that if your DOMDocument was loaded from HTML, where element and attribute names are case-insensitive, the DOM parser converts them all to lower-case, so your XPath queries will have to as well; '//A/@HREF' won't find anything even if the original HTML contained "<A HREF='example.com'>".
Please note that what clochix says is valid for *any* document which has a default namespace (as it is the case for XHTML).
This document :
<?xml version="1.0" encoding="UTF-8" ?>
<root xmlns="http://www.exemple.org/namespace">
<element id="1">
...
</element>
<element id="2">
...
</element>
</element>
must be accessed this way :
$document = new DOMDocument();
$document->load('document.xml');
$xpath = new DOMXPath($document);
$xpath->registerNameSpace('fakeprefix', 'http://www.exemple.org/namespace');
$elements = $xpath->query('//fakeprefix:element');
Of course, there is no prefix in the original document, but the DOMXPath class *needs* one, whatever it is, if you use a default namespace. It *doesn't work* if you specify an empty prefix like this :
$xpath->registerNameSpace('', 'http://www.exemple.org/namespace');
Hope this help to spare some time...
If you want to perform queries on XHTML documents, you must fix a default namespace:
$doc = new DOMDocument;
$doc->preserveWhiteSpace = true;
$doc->resolveExternals = true; // for character entities
$doc->load("http://www.w3.org/");
$xpath = new DOMXPath($doc);
// won't work
$entries = $xpath->query("//div");
// you should use :
$xpath->registerNamespace("html", "http://www.w3.org/1999/xhtml");
$entries = $xpath->query("//html:div");
tried finding a node by it's text content?
// Get all elements that equal the string "test"
$query = "//*[.='test']";
You can transform the result nodes into new DOMDocument objects this way:
<?php
$result = $xpath->query($query);
$resultNode = $result->item(0);
$newDom = new DOMDocument;
$newDom->appendChild($newDom->importNode($resultNode,1));
print "<pre>" . htmlspecialchars($newDom->saveXML()) . "</pre>";
?>
Two great XPath references follow.
XPath in Five Paragraphs (finally!):
http://www.rpbourret.com/xml/XPathIn5.htm
The w3c spec actually has a bunch of helpful examples:
http://www.w3.org/TR/xpath#location-paths
Diese Seite bei php.net