(PHP 5, PHP 7)
XSLTProcessor::transformToXml — Transform to XML
$doc
) : stringTransforms the source node to a string applying the stylesheet given by the xsltprocessor::importStylesheet() method.
The result of the transformation as a string or FALSE on error.
Example #1 Transforming to a string
<?php
// Load the XML source
$xml = new DOMDocument;
$xml->load('collection.xml');
$xsl = new DOMDocument;
$xsl->load('collection.xsl');
// Configure the transformer
$proc = new XSLTProcessor;
$proc->importStyleSheet($xsl); // attach the xsl rules
echo $proc->transformToXML($xml);
?>
以上例程会输出：
Hey! Welcome to Nicolas Eliaszewicz's sweet CD collection! <h1>Fight for your mind</h1><h2>by Ben Harper - 1995</h2><hr> <h1>Electric Ladyland</h1><h2>by Jimi Hendrix - 1997</h2><hr>
Note that the method's name is sort of deceiving, because it does not only output XML, but any string that is generated by the processor. It should rather be called transformToString. So if your output method is "text/plain", for example, this is the way to receive the resulting string.
Entities are omitted from the output with the code above.
The symptom was that
-- which should work with UTF-8 encoding --
did not even get to XSLTProcessor, let alone through it.
After too much hacking I discovered the simple fix:
set substituteEntities to true in the DOMDocument for the XSL file.
That is, replace the loading of the xsl document with
<?php
$xsl = new DOMDocument;
$xsl->substituteEntities = true; // <===added line
$xsl->load('collection.xsl');
?>
However, this fails when data entries have HTML entity references. (Some database entries may even contain user generated text.) libxml has the pedantic habit of throwing a FATAL error for any undefined entitiy. Solution: hide the entities so libxml doesn't see them.
<?php
function hideEntities($data) {
return str_replace("&", "&", $data);
}
?>
You could just add this to the example, but it is tidier to define a function to load data into a DOMDocument. This way you don't need entity declarations in catalog.xsl, either.
<?php
// Added function for Example #1
/** Load an XML file and create a DOMDocument.
Handles arbitrary entities, even undefined ones.
*/
function fileToDOMDoc($filename) {
$dom= new DOMDocument;
$xmldata = file_get_contents($filename);
$xmldata = str_replace("&", "&", $xmldata); // disguise &s going IN to loadXML()
$dom->substituteEntities = true; // collapse &s going OUT to transformToXML()
$dom->loadXML($xmldata);
return $dom;
}
// Compare with Example #1 Transforming to a string
// Load the XML sources
$xml = fileToDOMDoc('collection.xml');
$xsl = fileToDOMDoc('collection.xsl');
// Configure the transformer
$proc = new XSLTProcessor;
$proc->importStyleSheet($xsl);
// transform $xml according to the stylesheet $xsl
echo $proc->transformToXML($xml); // transform the data
?>
If you retrieve "false" from the transformToXML method, use libxml_get_last_error() or libxml_get_errors() to retrieve the errors.
To prevent your xsl file from automatically prepending
<?xml version="1.0"?>
whilst keeping the output as xml, which is preferable for a validated strict xhtml file, rather specify output as
<xsl:output method="xml" omit-xml-declaration="yes" />
If your xsl doesn't have <html> tags. The output will contain <?xml version="1.0"?>. To fix this you can add the following to your xsl stylesheet:
<xsl:output
method="html" />
How to fix:: *Fatal error: Call to undefined method domdocument::load()*
If you get this error, visit the php.ini file and try commenting out the following, like this:
;[PHP_DOMXML]
;extension=php_domxml.dll
Suddenly, the wonderfully simple example above works as advertised.
$domTranObj = $xslProcessor->transformToDoc($domXmlObj);
$domHtmlText = $domTranObj->saveHTML();
Do fix the <meta> for valid XHTML but do not correctly end empty node like <br /> which ouput like this : <br></br>
Some browser note this as 2 different <br /> ...
To fix this use
$domTranObj = $xslProcessor->transformToDoc($domXmlObj);
$domHtmlText = $domTranObj->saveXML();
I noticed an incompatibility between libxslt (php4) and the transformation through XSLTProcessor.
Php5 and the XSLTProcessor seem to add implicit CDATA-Section-Elements.
If you have an xslt like
<script type="text/javascript">
foo('<xsl:value-of select="bar" />');
</script>
It will result in
<script type="text/javascript"><![CDATA[
foo('xpath-result-of-bar');
]]></script>
(at least for output method="xml" in order to produce strict xhtml with xslt1)
That brings up an error (at least) in Firefox 1.5 as it is no valid javascript.
It should look like that:
<script type="text/javascript">//<![CDATA[
foo('xpath-result-of-bar');
]]></script>
As the CDATA-Section is implicit, I was not able to disable the output or to put a '//' before it.
I tried everything about xsl:text disable-output-escaping="yes"
I also tried to disable implicit adding of CDATA with <output cdata-section-elements="" />
(I thought that would exclude script-tags. It didn't).
The solution:
<xsl:text disable-output-escaping="yes"><script type="text/javascript">
foo('</xsl:text><xsl:value-of select="bar" /><xsl:text disable-output-escaping="yes">');
</script></xsl:text>
Simple, but it took me a while.
The transformToXML function can produce valid XHTML output - it honours the <xsl:output> element's attributes, which defines the format of the output document.
For instance, if you want valid XHTML 1.0 Strict output, you can provide the following attribute values for the <xsl:output> element in your XSL stylesheet:
<xsl:output
method="xml"
doctype-system="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"
doctype-public="-//W3C//DTD XHTML 1.0 Strict//EN" />
The function transformToXML has a problem with the meta content type tag. It outputs it like this:
<meta http-equiv="Content-Type" content="text/html; CHARSET=gb2312">
which is not correct X(HT)ML, because it closes with '>' instead of with '/>'.
A way to get the output correct is to use instead of transformToXML first transformToDoc anf then saveHTML:
$domTranObj = $xslProcessor->transformToDoc($domXmlObj);
$domHtmlText = $domTranObj->saveHTML();
transformToXML, if you have registered PHP functions previously, does indeed attempt to execute these functions when it finds them in a php:function() pseudo-XSL function. It even finds static functions within classes, for instance:
<xsl:value-of select="php:function('MyClass::MyFunction', string(@attr), string(.))" disable-output-escaping="yes"/>
However, in this situation transformToXML does not try to execute "MyClass::MyFunction()". Instead, it executes "myclass:myfunction()". In PHP, since classes and functions are (I think) case-insensitive, this causes no problems.
A problem arises when you are combining these features with the __autoload() feature. So, say I have MyClass.php which contains the MyFunction definition. Generally, if I call MyClass::MyFunction, PHP will pass "MyClass" to __autoload(), and __autoload() will open up "MyClass.php".
What we have just seen, however, means that transformToXML will pass "myclass" to __autoload(), not "MyClass", with the consequence that PHP will try to open "myclass.php", which doesn't exist, instead of "MyClass.php", which does. On case-insensitive operating systems, this is not significant, but on my RedHat server, it is--PHP will give a file not found error.
The only solution I have found is to edit the __autoload() function to look for class names which are used in my XSL files, and manually change them to the correct casing.
Another solution, obviously, is to use all-lowercase class and file names.