JSON to XML

What it does

JSON and XML are both trees and they disagree about almost everything else. There are no arrays in XML, no null, and no single required root — so a conversion is a set of conventions, and they should be visible.

Emit elements for object members, repeat the element for each array item, and write a declared root. Report the convention used for arrays, nulls and the root.

Why it works this way

An array becomes either repeated sibling elements or a wrapper element with the items inside. Both are common and they are not interchangeable — a reader expecting one gets nothing from the other.

JSON null has no XML form. An empty element, an explicit nil attribute and an omitted element are three different statements, and choosing silently means the reader cannot tell which was meant.

XML requires a single root. A JSON array at the top level has none, so one is supplied by name and the name is reported because it is a choice, not a fact.

Examples

{"a": [1, 2]}

<a>1</a><a>2</a>

{"a": null}

<a/>

Inputs and outputs

• JSON text
• array convention: repeated element / wrapped in a container
• null: empty element / xsi:nil / omit
• root element name

• XML text
• the convention applied, named

Edge cases

• arrays of arrays
• keys that are not valid XML names
• text containing < and &
• an array at the top level — needs a supplied root
• mixed arrays of objects and scalars

Sources

• W3C XML 1.0, section 2 (Documents must have exactly one root element)