JSON Schema Validator

Example

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "required": ["sku", "price"],
  "properties": {
    "sku": { "type": "string", "pattern": "^[A-Z]{3}-\\d{4}$" },
    "price": { "type": "number", "exclusiveMinimum": 0 },
    "tags": { "type": "array", "items": { "type": "string" } }
  }
}

{
  "sku": "abc-12",
  "price": 0,
  "tags": ["sale", 3]
}

The validator reports:

Note that "price": 0 fails exclusiveMinimum: 0 but would pass minimum: 0, and the pattern needs the backslash escaped in JSON (\\d).

JSON Schema versions

The $schema keyword at the top of a schema says which version it is written in:

The versions are mostly compatible. The changes that bite: definitions became $defs in 2019-09, array-form items became prefixItems in 2020-12, and draft 4’s boolean exclusiveMinimum: true became a number in draft 6.

Common mistakes

additionalProperties: false with allOf

additionalProperties only sees the properties next to it, not those in allOf branches, so combining schemas this way rejects every property. Use unevaluatedProperties: false (2019-09 and later) at the top level instead.

required inside a property

required is a list on the object, not a flag on each property: write "required": ["email"] next to properties, not "email": { "required": true } (that was draft 3).

Forgetting type

Keywords only apply to matching types. { "minimum": 1 } accepts the string "0", because minimum only checks numbers. Add "type": "integer" to reject it.

Validate in code and on the command line

// npm install ajv ajv-formats
import Ajv2020 from 'ajv/dist/2020.js';
import addFormats from 'ajv-formats';

const ajv = new Ajv2020({ allErrors: true });
addFormats(ajv);

const validate = ajv.compile(schema);
if (!validate(data)) {
  for (const e of validate.errors) console.log(e.instancePath || '/', e.message);
}

# pip install jsonschema
from jsonschema import Draft202012Validator

validator = Draft202012Validator(schema, format_checker=Draft202012Validator.FORMAT_CHECKER)
for error in sorted(validator.iter_errors(data), key=lambda e: e.path):
    print(error.json_path, error.message)

# check-jsonschema: JSON, YAML and TOML files, plus built-in schemas (pip install check-jsonschema)
check-jsonschema --schemafile user.schema.json user.json
check-jsonschema --builtin-schema vendor.github-workflows .github/workflows/ci.yml

# ajv-cli (npm install -g ajv-cli ajv-formats)
ajv validate --spec=draft2020 -c ajv-formats -s user.schema.json -d user.json

Don’t have a schema yet? Generate one from sample data with JSON to JSON Schema, then tighten it by hand. For syntax errors in the JSON itself, use the JSON Validator.

Frequently asked questions