Custom GraphQL types

Introduction

You can add custom GraphQL types in Hasura that you can utilise for defining your actions.

Limitations

It is currently not possible to define Interfaces and Union types as custom types

Object types

The most basic components of a GraphQL schema are object types, which just represent a kind of object a GraphQL query can return, and what fields it has. In the GraphQL SDL, we might represent it like this:

type UserInfo {
  accessToken: String!
  userId: Int!
}

This is an object type called UserInfo that has two fields:

  • accessToken: This field is of type String! (non-nullable String)
  • userId: This field is of type Int! (non-nullable Int)

See reference

Limitations

Hasura does not allow a field of an object type to be another object type, i.e. the fields of an object type can only be scalars and enums.

For a more complicated structure, the object types can be connected to the rest of the graph via relationships.

Relationships

Custom object types can be connected to the rest of the graph by setting up table relationships with tables/views.

For example, given the object type:

type UserInfo {
  accessToken: String!
  userId: Int!
}

and tables:

user (id int, name text)
order (id int, user_id int, ...)

We can create:

  • an object relationship called loggedInUser between the UserInfo object type and the user table via the UserInfo.userId and user.id fields.
  • an array relationship called userOrders between the UserInfo object type and the order table via the UserInfo.userId and order.user_id fields.

The object type will now be modified as:

type UserInfo {
  accessToken: String!
  userId: Int!
  loggedInUser: user
  userOrders: [order]
}

Note

Only fields with non-list scalar types (e.g. Int, String) can be used to define relationships

Input types

You can pass complex objects as arguments to queries and mutations. This is particularly valuable in cases where you might want to pass in a whole object to be created. In the GraphQL SDL, input types look exactly the same as regular object types, but with the keyword input instead of type:

input LoginInfo {
  username: String!
  password: String!
}

A field of an input type could be a scalar, an enum or another input type.

See reference

Scalar types

A GraphQL object type has a name and fields, but at some point those fields have to resolve to some concrete data. That’s where the scalar types come in: they represent the leaves of the query.

Inbuilt scalars

Hasura comes with some default GraphQL scalars that you can directly start using while defining your actions:

  • Int: A signed 32‐bit integer.
  • Float: A signed double-precision floating-point value.
  • String: A UTF‐8 character sequence.
  • Boolean: true or false.
  • ID: The ID scalar type represents a unique identifier, often used to refetch an object or as the key for a cache. The ID type is serialized in the same way as a String; however, defining it as an ID signifies that it is not intended to be human‐readable.

See reference

Custom scalars

Hasura allows you to define custom scalars. For example, if you want to define a scalar called Date, you can define it like.

scalar Date

These scalars can be used as arguments of queries and mutations or as fields of object types and input types.

Postgres scalars

Postgres base types are implicitly made available as GraphQL scalars; there is no need to declare them separately. For example, in the definition

type User {
  id: uuid!
  name: String!
  location: geography
}

the uuid and geography types are assumed to refer to Postgres scalars (assuming no other definition for them is provided).

Enum types

Enums are a special kind of scalar that is restricted to a particular set of allowed values. This allows you to:

  • Validate that any arguments of this type are one of the allowed values
  • Communicate through the type system that a field will always be one of a finite set of values

Here’s what an enum definition might look like in the GraphQL schema language:

enum Color {
  RED
  GREEN
  BLUE
}

This means that wherever we use the type Color in our schema, we expect it to be exactly one of RED, GREEN, or BLUE.

See reference