XML Attributes and Default Namespaces Explained

In XML, default namespaces apply exclusively to elements and are not inherited by attributes. While an unqualified element inherits the default namespace declared on itself or an ancestor, an unqualified attribute always belongs to no namespace at all. Understanding this distinction is essential for correctly querying, validating, and processing XML documents that utilize namespaces.

The Rule: Default Namespaces Apply Only to Elements

A default namespace is declared using the xmlns attribute without a prefix (for example, xmlns="http://example.com/ns"). According to the W3C XML Namespaces specification, this declaration sets the namespace for the element on which it appears and all of its descendant elements that do not explicitly declare their own namespace.

However, default namespaces do not apply to attributes. An attribute without a namespace prefix is never in the default namespace; it is in “no namespace” (also known as the null or unqualified namespace).

Example of Attribute Behavior

Consider the following XML snippet:

<person xmlns="http://example.com/ns" id="101" status="active">
    <name>Jane Doe</name>
</person>

In this example: * The <person> element is in the http://example.com/ns namespace. * The <name> element inherits the default namespace and is also in the http://example.com/ns namespace. * The id and status attributes are in no namespace, despite being declared directly on an element with a default namespace.

How to Place an Attribute in a Namespace

To assign an attribute to a specific namespace, you must explicitly use a prefixed namespace declaration.

<person xmlns="http://example.com/ns" 
        xmlns:meta="http://example.com/meta"
        id="101" 
        meta:status="active">
    <name>Jane Doe</name>
</person>

In this case: * id remains in no namespace. * meta:status is explicitly placed in the http://example.com/meta namespace via the meta prefix.

Why This Rule Exists

The W3C specification designed attribute scoping this way because attributes already derive their context directly from the element to which they belong. An attribute named id on a <person> element is inherently interpreted as the identifier of that specific person. Automatically putting attributes into the default namespace would create unnecessary complexity and ambiguity when integrating different schemas and parsing documents with technologies like XPath or XSLT.

Implications for XPath and Processing

When querying XML documents with tools like XPath, you must treat unprefixed attributes as having no namespace: