Can we avoid dataloaders at all?

[frederikhors]

Can we do the same as https://entgo.io/docs/tutorial-todo-gql-field-collection?

Basically it would be great to have the option to avoid dataloaders altogether.

[billy1624]

Hey @frederikhors, you're everywhere hahaaa :)

May I ask why dataloaders are a bad thing thus to be avoided?

CC @karatakis

[frederikhors]

Before knowing entgo (and entgql) I also thought the same thing: dataloaders are a good and right thing.

As soon as I started using entgql with the field-collection feature I immediately changed my mind because it is wonderful not to have to write a single extra line of code for the dataloaders.

The graphql query is read and the database is asked for exactly what it takes to satisfy the request.

Please, you who have managed to create all this, you who are geniuses, please also succeed in this: it would be wonderful!

[tyt2y3]

I'd love to hear your feedback as you play around more!

Terminology might be different, but Seaography works in a similar way to tackle the N+1 problem

If you try the example,

{
  film(pagination: { limit: 10, page: 0 }, orderBy: { title: ASC }) {
    data {
      title
      description
      releaseYear
      filmActor {
        actor {
          firstName
          lastName
        }
      }
    }
  }
}

Only 3 queries are executed: once from film, once from film_actor, once from actor.

(actually there is a 4th for pagination, but definitely not N+1)