JSON vs YAML vs XML vs CSV: Choosing a Data Format

All four formats store structured data as text. They differ in how much structure they support, how easy they are to read, and where people expect to find them.

JSON

The default for web APIs. It handles objects, lists, numbers, strings, true, false and null. It is strict, so mistakes are caught early, but it has no comments. Use the JSON Formatter to tidy and validate it.

YAML

Readable by humans, because it uses indentation instead of brackets, and it allows comments. That makes it popular for configuration files such as Docker Compose, GitHub Actions and Kubernetes. The catch: indentation mistakes silently change the meaning, and a few values like no or on can be read as booleans. Convert between the two with the YAML to JSON Converter.

XML

Older and more verbose, but good for documents with mixed text and markup, and still required by many enterprise systems, feeds and office file formats. It supports attributes and namespaces, which JSON doesn't. The XML to JSON Converter helps when you have to bridge the two.

CSV

Rows and columns, nothing more. It opens in any spreadsheet and is small and simple, which is why exports and imports use it. It can't hold nested data, and it has no types, so everything is text. Quoting rules vary, so a comma inside a value needs care. The CSV to JSON Converter turns rows into objects and back.

A quick guide

• Sending data between a browser and a server? JSON.
• Writing a config file people will edit by hand? YAML, or JSON if you want strictness.
• Working with an older system or a document format? XML.
• Moving a table in or out of a spreadsheet? CSV.

If you aren't sure, JSON is the safest choice because every language can read it.