SimpleXML has declaration of xmlns:xmlns="" - no way to remove

Question

So infuriating, I can hardly talk. I've assembled an RSS feed with SimpleXML and yet - it's using name spaces, which are right now. But, it's constantly trying to declare xmlns:xmlns="" in the root node, when output. Even though I do no such thing.

It starts with

$rssXML->addAttribute("version", '2.0');
$rssXML->addAttribute("xmlns:media", "http://search.yahoo.com/mrss/", '');
$rssXML->addAttribute("xmlns:dcterms", "http://purl.org/dc/terms/", '');

and after this I do:-

header("Content-Type: application/rss+xml");

echo $syndicationXML->asXML();

Yet it outputs :-

<?xml version="1.0"?>
<rss xmlns:xmlns="" version="2.0" xmlns:media="http://search.yahoo.com/mrss/" xmlns:dcterms="http://purl.org/dc/terms/"><channel>...

I do not understand all this namespace declaration. What's going on?

score 7 · Accepted Answer · edited May 02 '13 at 15:23

The problem with SimpleXML is that it's addAttribute function adds an attribute, not a namespace and although it seems like it does what you want, it's not meant to be used the way you are using it.

It's meant to add a value that's part of a particular namespace (specified as the third parameter), not to add the namespace itself. The reason why you end up with xmlns:xmlns is because SimpleXML found that you used the xmlns namespace when specifying the name xmlns:media for instance so it created an empty xmlns:xmlns.

Here are 2 solutions to your problem:

1. Specify in the namespaces in the constructor.

$rssXML = new SimpleXMLElement('<rss xmlns:media="http://search.yahoo.com/mrss/" xmlns:dcterms="http://purl.org/dc/terms/" />');
$rssXML->addAttribute('version', '2.0');

2. Replace xmlns:xmlns="" using preg_replace

echo preg_replace('/xmlns:xmlns=""\s?/', '', $rssXML->asXML());

Very clearly explained, very helpful and right. Thank you. :) — waxical, Jul 08 '11 at 16:44

score 1 · Answer 2 · edited May 23 '17 at 12:00

You wrote that you want to remove that. Best way is to not put it in the first place.

This works by prefixing with xmlns: (another time) - it somehow enables a special mode of operation in simplexml - and not providing an empty namespace-URI (otherwise you actually ask for adding that):

$rssXML = new SimpleXMLElement('<rss/>');
$rssXML->addAttribute("version", '2.0');
$rssXML->addAttribute("xmlns:xmlns:media", "http://search.yahoo.com/mrss/");
$rssXML->addAttribute("xmlns:xmlns:dcterms", "http://purl.org/dc/terms/");
$rssXML->asXML('php://output');

This creates the following output (beautified for your reading pleasure):

<?xml version="1.0"?>
<rss version="2.0" 
     xmlns:media="http://search.yahoo.com/mrss/" 
     xmlns:dcterms="http://purl.org/dc/terms/"/>

This "cheat" was originally hinted in "Unable to add Attribute with Namespace Prefix using PHP Simplexml" but was missing here.

It works because the added node is an attribute, rather than a namespace declaration, and "xmlns:" is part of the local part of the attribute name, rather than a namespace prefix. You can use any prefix, not just "xmlns", as it gets ignored. It doesn't matter when serializing, but makes a difference when working with the object (e.g. using the arguments to [`SimpleXMLElement->attributes()`](http://php.net/SimpleXMLElement.attributes) and `SimpleXMLElement->children()`, or the namespace arguments to [`SimpleXMLElement->addChild()`](http://php.net/SimpleXMLElement.addChild)). — outis, Sep 14 '16 at 07:23

score 1 · Answer 3 · edited May 23 '17 at 11:46

SimpleXMLElement can only add namespaces indirectly. Since you can only add element and attribute nodes, not namespace declarations, text nodes or other node types, you must somehow make use of SimpleXMLElement->addAttribute() or SimpleXMLElement->addChild(). The latter will only add new namespaces to the child, so it's of no use here. addAttribute() will add the namespace along with an attribute to the given element, so if you then remove the attribute, you're left with the namespace as desired.

There isn't an obvious method of removing an attribute, but the use of unset as shown in several answers to "Remove a child with a specific attribute, in SimpleXML for PHP" can be adapted to the task, using SimpleXMLElement->attributes() to get a reference to the attribute.

<?php
$rssXML = new SimpleXMLElement('<rss/>');
$rssXML->addAttribute("version", '2.0');
# add a dummy attribute to get the namespace
$rssXML->addAttribute("media:_", '', "http://search.yahoo.com/mrss/");
unset($rssXML->attributes('media', TRUE)[0]);
# and again
$rssXML->addAttribute("dcterms:_", '', "http://purl.org/dc/terms/");
unset($rssXML->attributes('dcterms', TRUE)[0]);

echo $rssXML->asXML(), "\n";

Result:

<?xml version="1.0"?>
<rss xmlns:media="http://search.yahoo.com/mrss/" xmlns:dcterms="http://purl.org/dc/terms/" version="2.0"/>

The advantage this has over adding the namespace declaration as an attribute with a fake namespace prepended is it's treated as a namespace, rather than an attribute that happens to have something that looks like a namespace prefix in its name.

var_export($rssXML->getDocNamespaces());

Result:

array (
  'media' => 'http://search.yahoo.com/mrss/',
  'dcterms' => 'http://purl.org/dc/terms/',
)

SimpleXML has declaration of xmlns:xmlns="" - no way to remove

3 Answers3