updateQueries by id (dataIdFromObject)

[deoqc]

Related to this, but instead of reading from cache, writing to cache.

---

Is it possible to update a mutation by some id (when using dataIdFromObject), instead of by query name (as is done now)?

Let me be more clear:

const query = gql`
  query {
    user {
      id
      friends {
        id
        name
        age
      }
    }
  }
`

const mutation = gql`
  mutation makingFriends($id: ID!) {
    letsMakeFriend(id: $id) {
      id
      name
      age
    }
  }

I know the user id (say id='123'), then I want to:

updateObject: {
  '123': (previousObject, { mutationResult }) => // do the logic
}

---

When using dataIdFromObject we prefer to reason in a object by id sense instead of query / paths. For more complicated and nested situations (more common than this example above) is much better to reason and much less error prone.

Let's compare these 2:

update by query name / path

updateQueries: {
  myChoosenQueryAmongMany: (previousQueryResult, { mutationResult }) => (
    deepmerge(
      previousQueryResult,
      {my: {very: {deep: {nested: {structure: [mutationResult.data.myMutation]}}}}}
    )
  )
}

update by id

updateObject: {
  myUniqueIdOfStruture: (previousObject, { mutationResult }) => ([
    ...previousObject,
    mutationResult.data.myMutation
  ])
}

---

Edit: most of the cases I will just return the parent object and refetch the entire updated child (and let the automatic update by id kick in). But on few important cases, the updated child is huge and I want to return only the relevant data.

[helfer]

@deoqc I think this could be quite useful in some edge cases. Would you be willing to make a PR for this and get it implemented?

[helfer]

We've decided to implement this kind of imperative API for dealing with store updates right after we implement some logic for refetching, and before we implement any declarative API.

Closing this issue for lack of activity.

[Clcll]

@deoqc Did you solve your issue?