Set JSON Schema Required Fields From Sample Records

A single record can make an optional key look mandatory. With strict mode off, the linked generator puts a key in required when it appears in every observed object at the same location. Use rows with and without optional values to give it a useful first draft.

Worked example

Inventory rows with an optional bin

[
  {"id": "P-001", "bin": "A-7"},
  {"id": "P-002"}
]

Steps

  1. Paste the sample array into the generator with strict mode off. The second record deliberately omits the bin field.
  2. In the generated schema, look under items. Keep id in required to enforce the inventory contract, and keep bin in properties so its value is still checked when supplied.
  3. Validate the array, then remove id from one record and validate again. The original array should pass; the version missing id should fail at that record.

What to check

  • Confirm items.required contains id and omits bin. If your application sends null for an empty bin, allow null explicitly or include that case when inferring the schema.
  • A generator only knows what it sees. If an API identifies a field as mandatory, keep it required even if one input sample was incomplete.

Infer a schema from the sample

References