Should I use dataloaders in relay.Node.resolve_nodes?

I've got a node type like this:

@strawberry.type(name="Todo")
class TodoType(BaseNodeType):
    content: str
    completed: bool
    created_at: datetime
    updated_at: datetime | None

    @classmethod
    def from_orm(cls, todo: Todo) -> Self:
        """Construct a node from an ORM instance."""
        return cls(
            id=todo.id,
            content=todo.content,
            completed=todo.completed,
            created_at=todo.created_at,
            updated_at=todo.updated_at,
        )

    @classmethod
    async def resolve_nodes(  # noqa: ANN206
        cls,
        *,
        info: Info,
        node_ids: Iterable[str],
        required: bool = False,  # noqa: ARG003
    ):
        todos = await info.context.loaders.todo_by_id.load_many(node_ids)
        return list(map(cls.from_orm, [todo for todo in todos if todo is not None]))

@strawberry.type(name="Todo")
class TodoType(BaseNodeType):
    content: str
    completed: bool
    created_at: datetime
    updated_at: datetime | None

    @classmethod
    def from_orm(cls, todo: Todo) -> Self:
        """Construct a node from an ORM instance."""
        return cls(
            id=todo.id,
            content=todo.content,
            completed=todo.completed,
            created_at=todo.created_at,
            updated_at=todo.updated_at,
        )

    @classmethod
    async def resolve_nodes(  # noqa: ANN206
        cls,
        *,
        info: Info,
        node_ids: Iterable[str],
        required: bool = False,  # noqa: ARG003
    ):
        todos = await info.context.loaders.todo_by_id.load_many(node_ids)
        return list(map(cls.from_orm, [todo for todo in todos if todo is not None]))

Should I be using dataloaders to resolve my nodes here? If so, should I use a separate dataloader when required=False and a separate dataloader when required=True? Or am I completely wrong and should I just use my repository layer directly here?

6 Replies

Unknown User•3mo ago

Message Not Public

Aryan Iyappan•3mo ago

Oh, right! Thank you! So my doubt is.. is it a common pattern to implement two separate dataloaders: one which raises an error and another one which does not?

Unknown User•3mo ago

Message Not Public

Aryan Iyappan•3mo ago

off the top of my head, I think it would be nice if we could somehow pass the required argument to the dataloader, something like:

dataloader.load_many(keys, required=True)

dataloader.load_many(keys, required=True)

so that the dataloader fetch function can handle this logic..

Unknown User•3mo ago

Message Not Public

Aryan Iyappan•3mo ago

Oh that's actually a brilliant idea!

Gaming

Programming

Should I use dataloaders in relay.Node.resolve_nodes?