Relay Cursor Connections (Pagination)

In this section, we continue the GraphQL example by explaining how to implement the Relay Cursor Connections Spec. If you’re not familiar with the Cursor Connections interface, read the following paragraphs that were taken from relay.dev:

In the query, the connection model provides a standard mechanism for slicing and paginating the result set.

In the response, the connection model provides a standard way of providing cursors, and a way of telling the client when more results are available.

An example of all four of those is the following query:

  1. {
  2. user {
  3. id
  4. name
  5. friends(first: 10, after: "opaqueCursor") {
  6. edges {
  7. cursor
  8. node {
  9. id
  10. name
  11. }
  12. }
  13. pageInfo {
  14. hasNextPage
  15. }
  16. }
  17. }
  18. }

Clone the code (optional)

The code for this tutorial is available under github.com/a8m/ent-graphql-example, and tagged (using Git) in each step. If you want to skip the basic setup and start with the initial version of the GraphQL server, you can clone the repository as follows:

  1. git clone git@github.com:a8m/ent-graphql-example.git
  2. cd ent-graphql-example
  3. go run ./cmd/todo/

Add Annotations To Schema

Ordering can be defined on any comparable field of Ent by annotating it with entgql.Annotation. Note that the given OrderField name must be uppercase and match its enum value in the GraphQL schema.

ent/schema/todo.go

  1. func (Todo) Fields() []ent.Field {
  2. return []ent.Field{
  3. field.Text("text").
  4. NotEmpty().
  5. Annotations(
  6. entgql.OrderField("TEXT"),
  7. ),
  8. field.Time("created_at").
  9. Default(time.Now).
  10. Immutable().
  11. Annotations(
  12. entgql.OrderField("CREATED_AT"),
  13. ),
  14. field.Enum("status").
  15. NamedValues(
  16. "InProgress", "IN_PROGRESS",
  17. "Completed", "COMPLETED",
  18. ).
  19. Default("IN_PROGRESS").
  20. Annotations(
  21. entgql.OrderField("STATUS"),
  22. ),
  23. field.Int("priority").
  24. Default(0).
  25. Annotations(
  26. entgql.OrderField("PRIORITY"),
  27. ),
  28. }
  29. }

Order By Multiple Fields

By default, the orderBy argument only accepts a single <T>Order value. To enable sorting by multiple fields, simply add the entgql.MultiOrder() annotation to desired schema.

ent/schema/todo.go

  1. func (Todo) Annotations() []schema.Annotation {
  2. return []schema.Annotation{
  3. entgql.MultiOrder(),
  4. }
  5. }

By adding this annotation to the Todo schema, the orderBy argument will be changed from TodoOrder to [TodoOrder!].

Order By Edge Count

Non-unique edges can be annotated with the OrderField annotation to enable sorting nodes based on the count of specific edge types.

ent/schema/todo/go

  1. func (Todo) Edges() []ent.Edge {
  2. return []ent.Edge{
  3. edge.To("children", Todo.Type).
  4. Annotations(
  5. entgql.RelayConnection(),
  6. entgql.OrderField("CHILDREN_COUNT"),
  7. ).
  8. From("parent").
  9. Unique(),
  10. }
  11. }

Relay Cursor Connections - 图1The naming convention for this ordering term is: UPPER(<edge-name>)_COUNT. For example, CHILDREN_COUNT or POSTS_COUNT. :::

Order By Edge Field

Unique edges can be annotated with the OrderField annotation to allow sorting nodes by their associated edge fields. For example, sorting posts by their author’s name, or ordering todos based on their parent’s priority. Note that in order to sort by an edge field, the field must be annotated with OrderField within the referenced type.

The naming convention for this ordering term is: UPPER(<edge-name>)_<edge-field>. For example, PARENT_PRIORITY.

ent/schema/todo.go

  1. // Fields returns todo fields.
  2. func (Todo) Fields() []ent.Field {
  3. return []ent.Field{
  4. // ...
  5. field.Int("priority").
  6. Default(0).
  7. Annotations(
  8. entgql.OrderField("PRIORITY"),
  9. ),
  10. }
  11. }
  12. // Edges returns todo edges.
  13. func (Todo) Edges() []ent.Edge {
  14. return []ent.Edge{
  15. edge.To("children", Todo.Type).
  16. From("parent").
  17. Annotations(
  18. entgql.OrderField("PARENT_PRIORITY"),
  19. ).
  20. Unique(),
  21. }
  22. }

Add Pagination Support For Query

  1. The next step for enabling pagination is to tell Ent that the Todo type is a Relay Connection.

ent/schema/todo.go

  1. func (Todo) Annotations() []schema.Annotation {
  2. return []schema.Annotation{
  3. entgql.RelayConnection(),
  4. entgql.QueryField(),
  5. entgql.Mutations(entgql.MutationCreate()),
  6. }
  7. }
  1. Then, run go generate . and you’ll notice that ent.resolvers.go was changed. Head over to the Todos resolver and update it to pass pagination arguments to .Paginate():

ent.resolvers.go

  1. func (r *queryResolver) Todos(ctx context.Context, after *ent.Cursor, first *int, before *ent.Cursor, last *int, orderBy *ent.TodoOrder) (*ent.TodoConnection, error) {
  2. return r.client.Todo.Query().
  3. Paginate(ctx, after, first, before, last,
  4. ent.WithTodoOrder(orderBy),
  5. )
  6. }

Relay Cursor Connections - 图2Relay Connection Configuration

The entgql.RelayConnection() function indicates that the node or edge should support pagination. Hence,the returned result is a Relay connection rather than a list of nodes ([T!]! => <T>Connection!).

Setting this annotation on schema T (reside in ent/schema), enables pagination for this node and therefore, Ent will generate all Relay types for this schema, such as: <T>Edge, <T>Connection, and PageInfo. For example:

  1. func (Todo) Annotations() []schema.Annotation {
  2. return []schema.Annotation{
  3. entgql.RelayConnection(),
  4. entgql.QueryField(),
  5. }
  6. }

Setting this annotation on an edge indicates that the GraphQL field for this edge should support nested pagination and the returned type is a Relay connection. For example:

  1. func (Todo) Edges() []ent.Edge {
  2. return []ent.Edge{
  3. edge.To("parent", Todo.Type).
  4. Unique().
  5. From("children").
  6. Annotations(entgql.RelayConnection()),
  7. }
  8. }

The generated GraphQL schema will be:

  1. -children: [Todo!]!
  2. +children(first: Int, last: Int, after: Cursor, before: Cursor): TodoConnection!

Pagination Usage

Now, we’re ready to test our new GraphQL resolvers. Let’s start with creating a few todo items by running this query multiple times (changing variables is optional):

  1. mutation CreateTodo($input: CreateTodoInput!) {
  2. createTodo(input: $input) {
  3. id
  4. text
  5. createdAt
  6. priority
  7. parent {
  8. id
  9. }
  10. }
  11. }
  12. # Query Variables: { "input": { "text": "Create GraphQL Example", "status": "IN_PROGRESS", "priority": 1 } }
  13. # Output: { "data": { "createTodo": { "id": "2", "text": "Create GraphQL Example", "createdAt": "2021-03-10T15:02:18+02:00", "priority": 1, "parent": null } } }

Then, we can query our todo list using the pagination API:

  1. query {
  2. todos(first: 3, orderBy: {direction: DESC, field: TEXT}) {
  3. edges {
  4. node {
  5. id
  6. text
  7. }
  8. cursor
  9. }
  10. }
  11. }
  12. # Output: { "data": { "todos": { "edges": [ { "node": { "id": "16", "text": "Create GraphQL Example" }, "cursor": "gqFpEKF2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" }, { "node": { "id": "15", "text": "Create GraphQL Example" }, "cursor": "gqFpD6F2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" }, { "node": { "id": "14", "text": "Create GraphQL Example" }, "cursor": "gqFpDqF2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" } ] } } }

We can also use the cursor we got in the query above to get all items that come after it.

  1. query {
  2. todos(first: 3, after:"gqFpEKF2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU", orderBy: {direction: DESC, field: TEXT}) {
  3. edges {
  4. node {
  5. id
  6. text
  7. }
  8. cursor
  9. }
  10. }
  11. }
  12. # Output: { "data": { "todos": { "edges": [ { "node": { "id": "15", "text": "Create GraphQL Example" }, "cursor": "gqFpD6F2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" }, { "node": { "id": "14", "text": "Create GraphQL Example" }, "cursor": "gqFpDqF2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" }, { "node": { "id": "13", "text": "Create GraphQL Example" }, "cursor": "gqFpDaF2tkNyZWF0ZSBHcmFwaFFMIEV4YW1wbGU" } ] } } }

Great! With a few simple changes, our application now supports pagination. Please continue to the next section where we explain how to implement GraphQL field collections and learn how Ent solves the “N+1 problem” in GraphQL resolvers.