Convert XML to JSON

How to use it#

1. Drop an .xml file on the box, or paste the XML into it.
2. Choose indented or minified JSON, and whether numbers and booleans become real JSON values.
3. Copy the JSON, or download it as a .json file named after the one you dropped.

How the XML maps onto JSON#

XML has two things JSON does not: attributes, and text that sits beside other content. Each needs a convention, and this page uses one of the most common. An attribute becomes a key with @ in front of its name, so id="7" becomes "@id": 7. Text inside an element that also has attributes or children goes under a key called #text. An element holding nothing but text is simply that text, which keeps the everyday case short.

The root element becomes the single top-level key, so a document that starts with <catalog> becomes an object with one key, catalog. Namespace prefixes are kept as part of the name, soap:Body stays soap:Body, and the xmlns declarations come through as ordinary @ attributes. So a SOAP response, an RSS or Atom feed, a sitemap.xml, or a Maven pom.xml converts like any other document.

The JSON to XML converter reads the same convention, so a document converted here and back comes out as the same XML, formatted. The exceptions are the things JSON has no room for, covered below.

Why an element is sometimes an array and sometimes not#

XML repeats an element to make a list, and JSON needs an array for that. When an element appears more than once under the same parent, its values become an array; when it appears once, it is a single value. That is the right reading of each document on its own, and it is also the most common surprise: an order with one line item gives an object where an order with two gives an array.

Code that reads the JSON should expect either. In JavaScript, [].concat(value) turns both forms into an array in one step. There is no way to tell from a single document which elements are meant to repeat, so no converter can get this right without a schema.

What is left behind#

Comments, processing instructions, and the DOCTYPE are dropped, since JSON has nowhere to put them. CDATA sections are read as the text they contain. Leading and trailing whitespace inside an element is trimmed, which is what almost every XML file indented for reading needs, but which does change a value whose spaces were deliberate.

Mixed content is the real limit. In <p>Hello <b>world</b>.</p> the text and the bold element are interleaved, and JSON objects have no order between different keys, so the result keeps the words and the element but not where the element sat in the sentence. XML used for data rarely does this; XHTML and document formats do all the time, and for those the XML formatter is the better tool.