Create Relationships

Now that we've created our subgraphs, we can start to model the relationships between them. While each of these subgraphs can function independently, we can also model relationships between them to create a single, unified API. This unlocks powerful queries and mutations that span across multiple data sources.

We'll use three examples to demonstrate how to create relationships between subgraphs:

• Users to Orders: connecting our app and fulfillment_services subgraphs.
• Orders to Products: connecting our fulfillment_services and product_management subgraphs.
• Users to Transactions: connecting our app and payment_processing subgraphs.

We can create these relationships using the same metadata structure we used to create relationships within a single subgraph. All we need to do is specify the subgraph name in the target field.

Create a relationship between users and orders

Open Users.hml and add the following relationship at the bottom of the file:

Copy
---
kind: Relationship
version: v1
definition:
  name: orders
  source: Users
  target:
    model:
      name: Orders
      relationshipType: Array
      subgraph: fulfillment_services
  mapping:
    - source:
        fieldPath:
          - fieldName: id
      target:
        modelField:
          - fieldName: customer_id

In this example, we're creating an array relationship between the Users and Orders models. This means that a single user can have multiple orders. We're also specifying the fulfillment_services subgraph as the target of this relationship.

Create a relationship between orders and products

Open Orders.hml and add the following relationship at the bottom of the file:

Copy
---
kind: Relationship
version: v1
definition:
  name: products
  source: Orders
  target:
    model:
      subgraph: product_management
      name: Products
      relationshipType: Array
  mapping:
    - source:
        fieldPath:
          - fieldName: productId
      target:
        modelField:
          - fieldName: id

In this example, we're creating an object relationship between the Orders and Products models. This means that a single order can have a single product. We're also specifying the product_management subgraph as the target of this relationship.

Create a relationship between users and transactions

Open Users.hml again and add the following relationship at the bottom of the file:

Copy
---
kind: Relationship
version: v1
definition:
  name: payments
  source: Users
  target:
    model:
      subgraph: payment_processing
      name: Transactions
      relationshipType: Array
  mapping:
    - source:
        fieldPath:
          - fieldName: id
      target:
        modelField:
          - fieldName: userId

In this example, we're creating an array relationship between the Users and Transactions models. This means that a single user can have multiple transactions. We're also specifying the payment_processing subgraph as the target of this relationship.

Test it out

We can now head to our project's console and see the visualization of our supergraph 🎉

We can also execute a query that spans across multiple subgraphs. For example, we can query for a user's orders, its associated products, and any payment transactions associated with the user:

Copy
query SupergraphQuery {
  app_users {
    id
    email
    orders {
      id
      orderDate
      products {
        id
        name
      }
    }
    payments {
      id
      amount
      transactionDate
      status
    }
  }
}