Sanity Attribute Limit: Do Duplicate Field Names in Different Documents Count Separately?

Great question about how Sanity counts unique attribute paths! Let me clarify this for you based on the official documentation on attribute limits.

In your example with two documents:

Page1: { title: 'Sanity', subtitle: 'Sanity' }
Page2: { title: 'Sanity', subtitle: 'Sanity' }

The answer is more than 3 - it would be approximately 4-6 attributes depending on the system fields.

Here's the key principle: Sanity counts unique paths, not individual field instances. According to the documentation, a path is counted based on its unique combination of attribute name and datatype. The documentation provides this example:

{
  "foo": [
    {
      "bar": ...,
      "baz": ...
    },
    {
      "bat": {
        "bar": ...
      }
    }
  ]
}

This structure contains six unique paths:

• foo (array)
• foo[] (object)
• foo[].bar (string)
• foo[].baz (string)
• foo[].bat (object)
• foo[].bat.bar (string)

For your specific example:

In your two documents, you'd have:

1. _id (system field)
2. _type (system field)
3. title (your field)
4. subtitle (your field)

Plus potentially other system fields like _createdAt, _updatedAt, etc.

The crucial point: Even though you have TWO documents with the same fields, each unique path is only counted once. Using the same field across 1 document or 1,000 documents doesn't increase your attribute count.

Important clarifications:

• Nested paths count separately: If your fields are nested under different objects (like Page1.content.title vs Page2.metadata.title), those ARE counted as separate paths
• Only actual content counts: Just defining fields in your schema doesn't increase the count until you actually store data in those fields
• Paths are reusable: The same path used multiple times across documents only counts once

If you're hitting your 4K limit:

You can check your current attribute count at: https://<projectId>.api.sanity.io/v1/data/stats/<datasetName> (look for fields.count.value)

Common causes of hitting the limit include:

• Deeply nested structures (especially with block content)
• Field-level translations/localization (creates multiplier effect)
• Many different document types with unique fields

Consider using document-level localization instead of field-level, and arrays instead of deeply nested objects for page builders to keep your attribute count manageable.