JSON to XML

XML is still what SOAP services, enterprise integrations, and document formats like DOCX and RSS expect. Converting JSON to it means resolving a genuine mismatch rather than reformatting: JSON has arrays and a real null, XML has neither, and XML has attributes, which JSON does not.

Four rules cover the whole conversion:

• []An array repeats the parent tag once per element. XML models lists by repetition, so there is nothing else to map it to.
• nullBecomes an empty element carrying xsi:nil="true", with the xmlns:xsi namespace declared on the root — and only when a null is actually present.
• 2faA key that is not a legal element name is sanitised: spaces become underscores, and a leading digit gets an underscore in front of it.
• <root>A top-level array is wrapped, with each element as <item>, because a document with several root elements is not well-formed.

Nothing here writes attributes. Every JSON key becomes a child element, which is verbose but unambiguous — the alternative is guessing which keys were "really" attributes, and a wrong guess is silent.

Why is my array not visible as an array?

XML has no array type. A list of three items is the same element three times over, which is the conventional encoding — but it means a one-element array and a plain value are indistinguishable in the output. Converting back gives you the scalar, not a list of one.

Is the xsi:nil output actually valid?

Yes. xsi is a namespace prefix, and using one without declaring it is a namespace error even though the document parses — so xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" goes on the root element whenever a null appears. Documents with no nulls do not carry a declaration they never use.

What happens to a key like "first name" or "2fa"?

They become first_name and _2fa. Element names cannot contain spaces or start with a digit, so there is no faithful option — and the rewrite is not reversible, which is worth knowing before treating this as a storage format rather than a hand-off.

Are the values escaped?

Text is escaped for &, <, and >, so a value containing markup stays a value. Those three are the ones that would break the document; quotes are only special inside attributes, and nothing here writes attributes.