XML Diff

Compare meaning rather than formatting

XML allows different spellings for equivalent content. Attribute order is not significant, entity references are decoded, and a CDATA section can contain the same text as an ordinary text node. Namespace prefixes are aliases: the namespace URI and local name identify an element or attribute.

<a:item xmlns:a="urn:shop" id="1" price="19.90">A &amp; B</a:item>
<!-- The following has the same expanded name, attributes and text. -->
<b:item xmlns:b="urn:shop" price="19.90" id="1"><![CDATA[A & B]]></b:item>

How paths and changes are reported

Paths use expanded names such as /{urn:shop}catalog/{urn:shop}item[1]/@price. The number identifies a child position. Identical subtrees align the sibling sequences so an inserted or removed node does not flag every later sibling as changed. Within unmatched sections, children are compared in order; identity is not guessed from ID attributes. Reordering remains significant. Removed and changed paths use original positions, while added paths use positions in the updated document. Downloaded JSON includes all differences, while the screen shows at most 300.

When to use a text diff instead

Use the text diff when comments, exact formatting or the original entity spelling matter. Use XML Formatter to check a document before comparing, or XPath Tester to inspect a particular subtree.

Frequently asked questions