My apologies if this is a redundant question but why do we need two schema.graphql files in the new graphcool api?

One is for your

server

and one is for your

database

.

I'm still a little bit confused to be honest. I take it the database one is used by graphcool to do the mapping then?

@rein did you read https://blog.graph.cool/graphql-databases-a-preview-into-the-future-of-graphcool-c1d4981383d9 and look into https://github.com/graphql-boilerplates/node-graphql-server? 🙂

@nilan Yes, I did. I might have not totally understood the separation though.

@rein I think I know what’s confusing you, the

database

schema is a type schema, which is only useful to

graphcool

to prepare database for you. Because you don’t need all of the endpoints

graphcool

features, you might want to strip down and extend some particular functionalities with your server. In order to tell the world what your app (server) can do, you have to define a new

app

schema. Your app and database have no direct relation, besides resolvers which might fetch data from the database. You might have also featured three schemas in case you would have decided to separate your data among two databases (in this case there would have been 2 database schemas and still only 1 app schema) I hope this make things a bit more clear 😄

Yes It does 🙂 thank you very much