Config Inference

Now that you've set up Postgres and added the "db" system, you can use config inference.

Add a db/2-user-roles.sql:

CREATE TYPE role AS ENUM ('editor', 'buyer', 'engineer', 'exec');
ALTER TABLE users ADD COLUMN roles role[];

And run docker compose down -v and docker compose up -d to pick up the new schema.

Now run rngo infer. This will output the following:

Updating the config:
 - adding property roles to stream users

and update .rngo/config to look something like this:

streams:
  users:
    # ...
    schema:
      type: object
      properties:
        id:
          type: integer
          minimum: 1
          rngo:
            cursor: ascending
        full_name:
          type: string
          rngo:
            enum: enum.fullName
        created_at:
          type: string
          format: date-time
          rngo:
            const: sim.now
        roles:
          type: array
          items:
            type: string
            enum:
              - editor
              - buyer
              - engineer
              - exec
      required:
        - id
        - full_name
        - created_at

Notice that it has preserved the customizations we made. It is safe to run rngo infer at any time - by default it is non-destructive and will only append to the schema. You should run in to bootstrap a new project and after subsequent schema updates.

Run rngo sim again - this time, the new roles column will have an array of the configured enum values, or null since it is not required like the other properties.