Prisma GraphQL APIReference

Mutations

Overview

How are mutations in the Prisma API generated?

The GraphQL API of a Prisma service is specified in the Prisma GraphQL schema. The Prisma GraphQL schema is auto-generated based on the service's datamodel:

The Mutation type of the Prisma GraphQL schema defines all the mutations the Prisma API accepts.

As an example, consider the following datamodel:

type User {
  id: ID! @unique
  name: String!
}

This is the Mutation type Prisma will generate:

type Mutation {
  createUser(data: UserCreateInput!): User!
  updateUser(data: UserUpdateInput!, where: UserWhereUniqueInput!): User
  deleteUser(where: UserWhereUniqueInput!): User
  upsertUser(
    where: UserWhereUniqueInput!
    create: UserCreateInput!
    update: UserUpdateInput!
  ): User!
  updateManyUsers(data: UserUpdateInput!, where: UserWhereInput!): BatchPayload!
  deleteManyUsers(where: UserWhereInput!): BatchPayload!
}

For every type in your datamodel, six mutation are generated. Taking above the above User type as an example, these mutations are:

  • createUser: Create a new User node
  • updateUser: Update an existing User node
  • deleteUser: Delete an existing User node
  • upsertUser: Update an existing User node; if it does not exist create a new User node
  • updateManyUsers: Update many User nodes at once
  • deleteManyUsers: Delete many User nodes at once

To inspect all the available operations of your Prisma API in detail, you can read the Prisma GraphQL schema of your Prisma service. It can be downloaded with the GraphQL CLI:

graphql get-schema --endpoint __YOUR_PRISMA_ENDPOINT__ --output prisma.graphql --no-all

Another way to learn about the concrete capabilities of your Prisma API is by exploring the auto-generated API documentation inside a GraphQL Playground. You can do so by clicking the green SCHEMA-button at the right edge of the Playground:

Datamodel for examples on this page

All example mutations on this page are based on a Prisma service configured with this datamodel:

type Post {
  id: ID! @unique
  title: String!
  published: Boolean! @default(value: "false")
  author: User
}

type User {
  id: ID! @unique
  age: Int
  email: String! @unique
  name: String!
  posts: [Post!]!
}

Mutation concepts

The Prisma API offers three kinds of mutations:

  • Object mutations: Create, update, upsert or delete a single node of a certain object type.
  • Nested mutations: Create, update, upsert, delete, connect, disconnect multiple nodes of two or more related types.
  • Batch mutations: Update and delete many nodes of a certain object type.

When working with the Prisma API, the following concepts are also useful to keep in mind:

  • Relation mutations: Connect and disconnect two nodes in a relation (this is done via nested mutations).
  • Input types: The Prisma API follows best practices of GraphQL mutation design and accepts mutation arguments as a single input object called data.

Object mutations

You can use object mutations to modify single nodes of a certain model.

Creating nodes

You can create new nodes using create mutations. These mutations take one input object called data wrapping all required fields of the respective type.

Create a new User node:

mutation {
  createUser(data: { age: 42, email: "alice@prisma.io", name: "Alice" }) {
    id
    name
  }
}

Updating nodes

You can change the values of one or more fields of a node using an update mutation. The node to be updated is selected using the where argument and the new values are provided via the data argument. Nodes can be selected by any field that's annotated with the @unique directive.

Change the name of a User node identified by its id:

mutation {
  updateUser(
    data: { name: "Alice" }
    where: { id: "cjcdi63j80adw0146z7r59bn5" }
  ) {
    id
    name
  }
}

Change the name of a User node identified by its email:

mutation {
  updateUser(data: { name: "Alice" }, where: { email: "alice@prisma.io" }) {
    id
    name
  }
}

Upserting nodes

When you want to either update an existing node, or create a new one in a single mutation, you can use upsert mutations.

Update the User with a certain email, or create a new User if a User with that email doesn't exist yet:

mutation {
  upsertUser(
    where: { email: "alice@prisma.io" }
    create: { email: "alice@prisma.io", age: 42, name: "Alice" }
    update: { age: 42 }
  ) {
    name
  }
}

Deleting nodes

To delete nodes, you need to select the node to be deleted via the where argument in a delete mutation.

Delete a User node by its id:

mutation {
  deleteUser(where: { id: "cjcdi63l20adx0146vg20j1ck" }) {
    id
    name
    email
  }
}

Delete a User node by its email:

mutation {
  deleteUser(where: { email: "cjcdi63l20adx0146vg20j1ck" }) {
    id
    name
    email
  }
}

Nested mutations

You can use create and update mutations to modify nodes across relations at the same time. This is referred to as nested mutations and is executed transactionally.

The MongoDB database connector currently does not support ACID transactions. Learn more in this GitHub issue.

Several nested mutation arguments exist:

  • create
  • update
  • upsert
  • delete
  • connect
  • disconnect

Rather than mapping out all possible scenarios at this point, we provide a list of examples.

It's recommended to explore the behaviour of different nested mutations by using the GraphQL Playground.

You can use the connect mutation within a nested input object field to connect to one or more related nodes.

Create a new Post node and connect it to an existing User node via the unique email field:

mutation {
  createPost(
    data: {
      title: "This is a draft"
      author: { connect: { email: "alice@prisma.io" } }
    }
  ) {
    id
    author {
      name
    }
  }
}

If you provide a create argument instead of connect within author, you would create a new User node and at the same time connect the new Post node to it:

mutation {
  createPost(
    data: {
      title: "This is a draft"
      author: { create: { name: "Bob", email: "bob@prisma.io", age: 42 } }
    }
  ) {
    id
    author {
      name
    }
  }
}

When using the createUser instead of a createPost mutation, you can actually create and connect to multiple Post nodes at the same time, because User has a to-many relation Post.

Create a new User node and directly connect it to several newly created Post nodes:

mutation {
  createUser(
    data: {
      name: "Bob"
      email: "bob@prisma.io"
      age: 42
      posts: {
        create: [
          { published: true, title: "Hello World" }
          { title: "GraphQL is great" }
        ]
      }
    }
  ) {
    id
    posts {
      id
    }
  }
}

Create a new User node and directly connect it to several newly created and existing Post nodes:

mutation {
  createUser(
    data: {
      name: "Bob"
      email: "bob@prisma.io"
      age: 42
      posts: {
        create: [
          { published: true, title: "Hello World" }
          { title: "GraphQL is great" }
        ]
        connect: [
          { id: "cjcdi63j80adw0146z7r59bn5" }
          { id: "cjcdi63l80ady014658ud1u02" }
        ]
      }
    }
  ) {
    id
    posts {
      id
    }
  }
}

When updating nodes, you can update one or more related nodes at the same time.

Update the title of a Post node that belongs to a certain User node:

mutation {
  updateUser(
    data: {
      posts: {
        update: [
          {
            where: { id: "cjcf1cj0r017z014605713ym0" }
            data: { title: "Hello World" }
          }
        ]
      }
    }
    where: { id: "cjcf1cj0c017y01461c6enbfe" }
  ) {
    id
    posts {
      id
    }
  }
}

update accepts a list of objects with where and data fields suitable for the updatePost mutation.

Update the name of an existing User node or create a new User node that belongs to a certain Post node:

Nested upserting works similarly:

mutation {
  updatePost(
    where: { id: "cjcf1cj0r017z014605713ym0" }
    data: {
      author: {
        upsert: {
          where: { id: "cjcf1cj0c017y01461c6enbfe" }
          update: { id: "cjcdi63l80ady014658ud1u02", name: "Alice" }
          create: { email: "alice@prisma.io", name: "Alice", age: 42 }
        }
      }
    }
  ) {
    id
    author {
      id
    }
  }
}

Disconnecting a Post node from a User node:

mutation {
  updateUser(
    where: { id: "cjcf1cj0c017y01461c6enbfe" }
    data: { posts: { disconnect: [{ id: "cjcdi63l80ady014658ud1u02" }] } }
  ) {
    id
    posts {
      id
    }
  }
}

When updating nodes, you can delete one or more related nodes at the same time.

Delete Post nodes belonging to a certain User node:

mutation {
  updateUser(
    data: {
      posts: {
        delete: [
          { id: "cjcf1cj0u01800146jii8h8ch" }
          { id: "cjcf1cj0u01810146m84cnt34" }
        ]
      }
    }
    where: { id: "cjcf1cj0c017y01461c6enbfe" }
  ) {
    id
  }
}

Scalar list mutations

When an object type has a field that is has a scalar list as its type, there are a number of special mutations available.

In the following datamodel, the User type has three such fields:

type User {
  id: ID! @unique
  scores: [Int!]! # scalar list for integers
  friends: [String!]! # scalar list for strings
  coinFlips: [Boolean!]! # scalar list for booleans
}

Creating nodes

When creating a new node of type User, a list of values can be provided for each scalar list field using set.

Create a new User and set the values for scores, friends and coinFlips:

mutation {
  createUser(data: {
    scores: { set: [1, 2, 3] }
    friends: { set: ["Sarah", "Jane"] }
    throws: { set: [false, false] }
  }) {
    id
  }
}

Updating nodes

When updating an existing node of type User, a number of additional operations can be performed on the scalar list fields (such as scores: [Int!]!):

  • set: Override the existing list with an entirely new list.
  • push (coming soon): Add one or more elements anywhere in the list.
  • pop (coming soon): Remove one or more elements from the beginning or the end of the list.
  • remove (coming soon): Remove all elements from the list that match a given filter.

push, pop and remove are not yet implemented. If you're curios what these are going to look like, you can get a preview in the respective specification.

Each scalar list field takes an object with a set field in an update mutation. The value of that field is a single value or a list of the corresponding scalar type.

Set the scores of an existing User node to [1]:

mutation {
  updateUser(
    where: { id: "cjd4lfdyww0h00144zst9alur" }
    data: { scores: { set: 1 } }
  ) {
    id
  }
}

Set the scores of an existing User node to [10,20,30]:

mutation {
  updateUser(
    where: { id: "cjd4lfdyww0h00144zst9alur" }
    data: { scores: { set: [10, 20, 30] } }
  ) {
    id
  }
}

Batch Mutations

Batch mutations are useful to update or delete many nodes at once. The returned data only contains the count of affected nodes.

Batch mutations are not triggering subscription events!

For updating many nodes, you can select the affected nodes using the where argument, while you specify the new values with data. All nodes will be updated to the same value.

Set the published fields of all unpublished Post nodes that were created in 2017 to true:

mutation {
  updateManyPosts(
    where: { createdAt_gte: "2017", createdAt_lt: "2018", published: false }
    data: { published: true }
  ) {
    count
  }
}

Delete all unpublished Post nodes where the name of the author is Alice:

mutation {
  deleteManyPosts(where: { published: false, author: { name: "Alice" } }) {
    count
  }
}