Skip to main content

Creating a New Schema

Workflow for generating a new schema and its CRUD resolvers. Read the code generation ordering gotcha first; it will save you a failed first run.

Create a New Schema File​

task db:newschema -- MyNewSchema

Update the Templated Schema​

The file will be located in internal/ent/schema/, the file name matching the schema name. Work through the template:

  • Keep the IDMixin and the standard annotations included in the template; removing IDMixin causes ID type conflicts with the rest of the graph
  • Add fields to the schema
  • Add edges for 1:1, 1:m, or m:m relationships with other objects
  • Add indexes where needed
  • Review and update the annotations on the templated schema
  • Add hooks for mutations and interceptors for queries; see Hooks and Interceptors
  • Review and update the privacy policy on the templated schema
Example

A common index pattern in our codebase allows soft-deleted names to be reused even when the name must be unique:

index.Fields("name").
    Unique().Annotations(entsql.IndexWhere("deleted_at is NULL")),

Generate the CRUD Resolvers​

task generate

This will take a few minutes the first time. The main tools that run:

ToolGenerates
entcORM code for the schema: field getters and setters, create/update/delete operations against the database
gqlgenGraphQL resolvers and basic query generation
gqlgencThe Go API client, from the queries in internal/graphapi/query
gencmdBase CRUD operations for CLI commands

Generate the Migrations​

Generate the goose and atlas migrations based on the ent schema changes:

task db:create

Stub Out the CRUD CLI Commands​

task cli:generate
info

The cli flags for create and update, along with field output (in root.go), will need to be manually updated