JSON Schema: Required, Optional and Nullable Fields

required controls presence, not value type

In an object schema, required lists property names that must exist. It does not automatically say whether a present property's value may be null.

{
  "type": "object",
  "required": ["name"],
  "properties": {
    "name": {"type": "string"}
  }
}

An optional property can still reject null

If a property is omitted from required, it may be absent. When it is present, it must still satisfy its declared schema. A property whose only allowed type is string therefore rejects null even when optional.

Allow null explicitly when the contract permits it

Modern JSON Schema can express a union of allowed types. If a field can be either a string or null, declare both instead of relying on an application-specific assumption.

{
  "properties": {
    "middleName": {"type": ["string", "null"]}
  }
}

Required and nullable can be combined

A property can be required to exist while allowing null as its value. That differs from an optional non-nullable property, which may be absent but must have a valid non-null value when present.

Empty values need their own constraints

An empty string is still a string, an empty array is still an array and an empty object is still an object. If those empty values are invalid, use constraints such as minLength, minItems or minProperties instead of treating them as null.

Test the complete state matrix

For important properties, write examples covering missing, present-valid, present-null, present-empty and wrong-type cases. This makes the intended contract obvious and prevents later changes from collapsing distinct states.

• Property missing.
• Property present with valid value.
• Property present with null.
• Property present with an empty value.
• Property present with the wrong type.