GraphQL API

Application of scopes

The access to resources is granted by the same scopes that are used on the REST API endpoints. We do not document them here in detail, but we give some guideline on how to find the scope best suitable for a use case in general:

For example, the manage_products scope is required for updating a Product with the Mutation.updateProduct service, the view_products scope would not be sufficient for this.

The view_published_products scope can be used by the Query.products service to retrieve data in masterData.current, but not in masterData.staged. The same applies for the Query.productProjectionSearch service that only returns data with this scope when the staged parameter is set to false in such queries.

Query GraphQL

You can access the GraphQL endpoint with following URL:

https://api.{region}.commercetools.com/{projectKey}/graphql

The endpoint accepts HTTP POST requests with following fields in a JSON body:

• query - String - GraphQL query as a string
• variables - Object - Optional - containing JSON object that defines variables for your query
• operationName - String - Optional - the name of the operation, in case you defined several of them in the query

Find below an example for a query for id and version of a Product with sku equal to SKU-123 to be executed as cURL command:

$ curl -X POST https://api.{region}.commercetools.com/my-shop/graphql \
  -H "Content-Type:application/json" \
  -H "Authorization:Bearer ..." \
  -d '{"query": "query ($sku: String!) {product(sku: $sku) {id version}}", "variables": {"sku": "SKU-123"}}'

Product attributes and custom fields

Product attributes and custom fields are project-specific dynamic fields that can be accessed using raw GraphQL fields. Raw GraphQL fields provide the values as raw JSON that you have to parse yourself. Queries for such fields ensure consistency and good performance.

Raw product attributes

Pseudo-example for using raw product attributes:

fragment variantFields on ProductVariant {
  sku
  attributesRaw {
    name
    value
  }
}

Raw custom fields

Pseudo-example for using raw custom fields:

fragment customFields on ProducePrice {
  custom {
    customFieldsRaw {
      name
      value
    }
  }
}

Reference Expansion

The GraphQL API supports Reference Expansion, just like the HTTP API.

By convention, the GraphQL API offers two fields for each expandable reference:

• <fieldName>Ref - Fetches the Reference to the resource only (lower query complexity).
• <fieldName> - Fetches the expanded resource that is referenced (higher query complexity).
  Returns null if the referenced resource is not found.

Expanding a reference in the GraphQL API impacts the performance of the request, and adds to the complexity score. If you don't need the expanded reference for your use case, you should use the <fieldName>Ref to get better performance on your query.

For an example, if you want to obtain the number of child categories for a specific category, as well as identifiers for its ancestors, the following query would work:

{
  category(id: "some-uuid") {
    children {
      id
    }
    ancestors {
      id
    }
  }
}

The total number of child categories will be the size of children array, and the ancestors identifiers are also available from the ancestors. However, the complexity of the example above can be reduced. If you only need the number of child categories, the childCount field is the way to go. Also, in case you only need the identifiers for the ancestors, it's better to use the ancestorsRef field. This is how the less complex query looks like:

{
  category(id: "some-uuid") {
    childCount
    ancestorsRef {
      id
    }
  }
}

Query complexity

You can fetch a lot of useful information in a single HTTP request using GraphQL. This is really helpful to avoid parsing unused fields, and remove the unnecessary network overhead by reducing the number of requests. At the same time, it is important to remember that a single query can potentially generate a lot of database operations. Hence, you cannot assume that the response time for a query increases linear with the number of fields in the query.
The GraphQL schema defines a "cost" per field, which you can take advantage of by analyzing your queries, using following options:

• Inspecting the x-graphql-query-complexity response header and its value.
• Using GraphiQL's Profiling functionality to measure the impact.

To prevent overly complex queries from having negative impact on the platform, we block queries that exceed the complexity limit of 20 000. If the complexity score for your query is equal or higher than the limit, a QueryComplexityLimitExceeded error code will be returned.

Interactive GraphQL console

To explore the GraphQL API, you can use an interactive GraphiQL environment which is available as a part of our ImpEx & API Playground.

Below animation demonstrates how you can use this tool for autocompletion and to explore the documentation that is part of the GraphQL schema:

Beta functionalities BETA

Find below a list of functionality that is currently in beta.

• createdBy on Versioned
• lastModifiedBy on Versioned
• everything related to Store
• everything related to Extension
• FixedPriceDiscountValueInput on CartDiscount
• MultiBuyLineItemsTarget and MultiBuyCustomLineItemsTarget on CartDiscount
• everything related to MyCart, MyOrder, MyPayment, MyProfile, MyShoppingList
• Order Edits on Order
• NestedTypes on ProductType
• Subrate on TaxCategory
• storeProjection and localeProjection on productProjectionSearch