howtographql | The Fullstack Tutorial for GraphQL | GraphQL library

It is the Extractor pattern, you can reach the same result implementing the unapply method on your arbitrary object (like shown in the example). When you create a case class the compiler produces an unapply method for you, so you can do:

That's because the server has been written with the business rule in mind that a Post will always belong to a User. The database has a NOT NULL postedById field on the Link table i.e. a post will always have a user id attached to it. You need to make postedById field nullable in the Link model in the Prisma ORM schema. To fix this, make the following changes on server side code and relaunch the server

1. In the server folder go to schema.prisma file, and make both fields postedBy and postedById optional/nullable by suffixing them with ?

The first thing you should do is read up on Promises. Promises are a way in JavaScript to encapsulate computations that are still ongoing. This is usually the case when you talk to an external service like a database or the operating system. They have been replacing callback style APIs.

In GraphQL a resolver can either return a value or a Promise that resolves to a value. This means, you can freely choose returning a value or a Promise, but if you call a database function like Prisma, you will get a Promise back, so you are kind of forced to stay "in Promise land", as there is no way to turn a Promise into a value. You can only chain functions, that should be executed with the value "in the future" (with then).

The last concept to understand is async/await. These async syntax is an addition to JavaScript syntax, that makes working with Promises easier. With await, you can stop the execution of a function until a value in a Promise arrives. Now, this looks like you are turning a Promise back into a value, but in reality, you function implicitly returns a Promise. For the VM to know about this, you have to state, that a function might use async by adding the keyword await in front of the function.

So when do you use async for a resolver? You could do it all the time, and the code would be correct. But doing it, even when you don't need to (e.g. you are not talking to a service) might have some performance implications. So it's better to only do it, if you really want to use the await keyword somewhere. I hope this can get you started with the concepts above, there is really a lot to learn. Maybe just go with your intuition and TypeScript errors until you deeply understand what is going on.

This part is not strictly related to auth [reasons] ... creating post mutation will work without that. But:

• post mutation creates an [Link] entry with a relation (to User in prisma schema meaning docs );
• postedBy is a virtual field - 'Relation fields define connections between models at the Prisma level and do not exist in the database. ';
• post mutation returns a Post type - then 'asked' return object can contain postedBy field (and its User object/type subfield[-s]);
• when postedBy field is queried (as part of mutation result tree), it (related User object/type - graphql context) must be resolved using function postedBy(parent, args, context);

In resolver body we're using/working with "Prisma client" - ORM lib, then:

got solved the issue,if you face similar kind of issue refer:https://github.com/howtographql/react-apollo/issues/85#issuecomment-671262863

I think it depends on your use cases. Says you only maintain 1-way pointer that keeps books array in author document, if you have a use case that requires searching for author by a certain book, you will need to do a full collection scan of author collection to find all matches. So without much knowledge of your actual scenario, it could be hard for us to comment on the necessity of 2 way pointer in your database.

This document is a good piece for your reading.

It seems that your issue is related to this. Currently there's an issue with Node.js version 12.19.0 that Heroku uses and there's a dependency named undici that's causing the issue. It should work if you try it with Node 14 in Heroku.

Actually it helped to restart the GraphQL server once again via yarn start.

It not, try to rebuild all modules after cleaning.