Docs Menu

Docs HomeAtlas App Services

Schemas

On this page

  • Overview
  • What is a Schema?
  • Why Define a Schema?
  • Define a Schema
  • How App Services Enforces Schemas
  • App Services Schema vs Built-In Schema Validation

A schema is a JSON object that defines the structure and contents of your data. You can use Atlas App Services' BSON schemas, which extend the JSON Schema standard, to define your application's data model and validate documents whenever they're created, changed, or deleted.

Schemas represent types of data rather than specific values. App Services supports many built-in schema types. These include primitives, like strings and numbers, as well as structural types, like objects and arrays, which you can combine to create schemas that represent custom object types.

For example, this is a basic schema for data about cars and some car objects that conform to the schema:

Schemas are the specification for your application's data model. Once you've defined a schema, App Services provides you with additional tools and services to work with data that conforms to the schema.

App Services uses schemas in many application services:

  • Atlas Device Sync uses schemas to sync data between realms and MongoDB Atlas. App Services can also generate idiomatic SDK object models for you based on your schemas.

  • Data Access Rules validate that data conforms to your schema before and after every request. If any document fails validation, App Services prevents or rolls back the entire request.

A root-level collection schema can contain additional schemas that describe the type's properties. Each root-level schema is an object schema that has the following form:

{
"bsonType": "object",
"title": "<Type Name>",
"required": ["<Required Field Name>", ...],
"properties": {
"<Field Name>": <Schema>
}
}

You can use any of the supported schema types to configure the object's properties:

Note

To learn how to configure and deploy a schema in your app, see Enforce a Schema.

App Services validates all write operations (inserts, updates, and deletes) on a MongoDB collection against its collection schema. It checks every document before and after every request to ensure that all properties conform to the schema and that no invalid changes occured.

App Services evaluates the result of all document writes and compares them against the schema before committing the writes to your cluster. If the result of any write operation in a request does not match the schema, App Services returns an error to the user without applying any changes in the request.

Example

A collection has the following schema:

{
"title": "person",
"properties": {
"_id": { "bsonType": "objectId" },
"name": { "bsonType": "string" }
}
}

A user with permission to read and write all fields wants to update the name field of a particular document. They issue the following query:

collection.updateOne(
{ "_id": BSON.ObjectId("5ae782e48f25b9dc5c51c4d0") },
{ "$set": { "name": 42 } }
)

The query attempts to set the value of name to the number 42, but the schema requires the value to be a string. App Services will reject this write operation even though the user had permission to update the document because the write result does not conform to the schema.

A schema in App Services is not the same as MongoDB's built-in schema validation. Both use the JSON schema standard with additional support for BSON types. However, App Services does not use your cluster's built-in schema and may interact with your cluster in a way that is incompatible with a built-in schema.

If you want to use App Services schemas and your cluster's built-in schema validation at the same time, consider the following:

  • Set your cluster's schema validation level to "warn" initially. Then, monitor activity and address existing warnings. Once you're comfortable that both schema validation layers are compatible, you can upgrade the validation level to "error".

  • If you're using Device Sync, avoid required fields for embedded documents and arrays of embedded documents. The sync protocol may break a valid embedded object write into multiple equivalent writes that do not individually include all required fields.

  • If you're using Device Sync, avoid distinguishing between undefined, null, empty arrays, and embedded objects with no fields. The sync protocol treats these values as functional equivalents.

If you need help in working with both schema validation layers simultaneously, contact MongoDB Support.

←  Define a Data ModelEnforce a Schema →