Validate unique boolean field across documents in Sanity schema

You can achieve this with a custom validation rule that queries your dataset to check if any other document has isFeatured set to true. Here's how to implement it:

defineField({
  name: 'isFeatured',
  type: 'boolean',
  title: 'Featured',
  validation: (Rule) =>
    Rule.custom(async (value, context) => {
      // Only validate if isFeatured is being set to true
      if (!value) return true

      // Get the Sanity client
      const client = context.getClient({ apiVersion: '2023-01-01' })
      
      // Query for other documents with isFeatured = true
      // Exclude the current document using its _id
      const query = `*[_type == $type && isFeatured == true && _id != $id][0...2] {
        _id,
        title
      }`
      
      const params = {
        type: context.document?._type,
        id: context.document?._id || ''
      }
      
      const existingFeatured = await client.fetch(query, params)
      
      // If another document is already featured, show an error
      if (existingFeatured && existingFeatured.length > 0) {
        const doc = existingFeatured[0]
        return `Only one document can be featured at a time. "${doc.title || doc._id}" is already featured.`
      }
      
      return true
    })
})

Key points about this implementation:

1. getClient() access: The validation context provides access to the Sanity client via context.getClient(), which allows you to query your dataset during validation.

2. Excluding current document: The query uses _id != $id to exclude the current document, so you can save a document that already has isFeatured set to true without triggering a validation error.

3. Informative error message: The error message tells the editor which document is already featured, making it easy to find and update if needed.

4. Only validates when true: The validation only runs when isFeatured is true, so it doesn't unnecessarily query the database for every document save.

Important considerations:

• This validation only runs client-side in the Studio. If you're updating documents via the API, you'll need to implement the same logic server-side.
• Make sure to adjust the _type and field names to match your schema.
• Replace title in the query with whatever field you want to display in the error message (or just use _id if you don't have a title field).

This approach uses custom validation with async operations, which is perfect for cases where you need to validate against other documents in your dataset.