Hi guys! I need to implement validations on mutations in Prisma. The doc suggests to check values directly in the resolver, which is ok for me, but what about the checks on the relations? For example, take

Post

and

Tag

, let's say that in an update mutation on

Post

I need to create a

Tag

, how can validate the new tag?

what stops you from validating the tag?

hi nilan, in the

Post

mutation, the

Tag

(beign a relation) is not a plain object, but has the input structure like

connect: {id: "xyz"}

, same for create, disconnect and so on. So it's not as straightforward as validating the main object. Moreover I have nested relations, so I cannot check them one by one in this way. I would expect the inner relation to call my

Tag

mutation resolver and do there all the validations, but instead it is called the main mutation

context.db.mutation.createTag

, so I can't do checks in the middle

You can define the input fields yourself, so instead of

connect: {id: "xyz"}

you could define an input field

tagId: "xyz"

which you then can validate in your own resolver

yes, it could be a viable solution, even if we will lose a bit the advantage of using Prisma and its code generation. Not asking for ETA, but you know if someone is actively working on the issue 728?