CODEWITHSUNDEEP
graphqlgraphiqlfaunadb
Nick
Nick

Reputation: 73

FaunaDB GraphiQL will not allow me to update my schema when I specify a new input type

This is an example of a document I am trying to retrieve:

  {
  "ref": Ref(Collection("Word"), "270608756095582738"),
  "ts": 1594331477980000,
  "data": {
    "word": "ablatitious",
    "letters": [
      {
        "letter": "a",
        "occurrence": 2
      },
      {
        "letter": "b",
        "occurrence": 1
      },
      {
        "letter": "l",
        "occurrence": 1
      },
      {
        "letter": "t",
        "occurrence": 2
      },
      {
        "letter": "i",
        "occurrence": 2
      },
      {
        "letter": "o",
        "occurrence": 1
      },
      {
        "letter": "u",
        "occurrence": 1
      },
      {
        "letter": "s",
        "occurrence": 1
      }
    ],
    "length": 11
  }
}

This is my schema

 type Letter @embedded {
  letter: String! @unique
  occurrence: Int!
}

type Word {
  word: String! @unique
  letters: [Letter]!
  length: Int!
}

input LetterInput {
  letter: String!
  occurrence: Int!
} 

type Query {
  Word(length: Int!): Word
  WordByLetters(letters: [LetterInput!]): Word
}

This is the error I get when I attempt to update my schema with this schema:

Schema does not pass validation. Violations:

Type mismatch: field 'letters' defined at object 'Word' has type 'Letter'. (line 19, column 17):
  WordByLetters(letters: LetterInput): Word
                ^

If I switch the type LetterInput to Letter in the WordByLetters query type, I get this error

Type 'Letter' is not an input type type. (line 19, column 26):
  WordByLetters(letters: Letter): Word
                         ^

So, clearly, I need to use an input type, which makes sense. What does not make sense is the first of the two errors. Can somebody please explain why?

Upvotes: 0

Views: 218

Answers (1)

Brecht De Rooms
Brecht De Rooms

Reputation: 1822

FaunaDB dev advocate here.

I'm not an expert on our GraphQL (I'll poke one of the experts if you are still stuck). When you define queries like that in FaunaDB's GraphQL it will try to match the attribute (letters: Letter) to one of the attributes in the return type (Word).

Since that return type is Word

type Word {
  word: String! @unique
  letters: [Letter]!
  length: Int!
}

it's logical that you try to place Letter there. However if I'm not mistaking it should be the more exact which meant it has to be [ Letter ]. However, since Letter is a type, you will have to present an ID instead of the actual Letter object. You might try to define:

WordByLetters(letters: [ ID ] ): Word

instead but I'm honestly not sure whether you can place arrays in the arguments. And since what you need is actually this:

WordByLetters(letters: ID ): Word

which we don't really support atm if I'm not mistaking. What you could do however is to not define the letters as embedded, connect to the correct letters (instead of embedding them each time). You would then also change your schema slightly to have a many to many relation:

type Letter {
  letter: String! @unique
  occurrence: Int!
  words: [ Word! ]!
}

You could predefine all letters to get easy access to the IDs of the letters. If you do that (I don't think this would work with embedded but I'm not 100% certain) you could find the words via the letters:

type Query {
  findLetterByID(id: ID!): Letter
}

which is a query that is already defined by FaunaDB (at least if Letters would not be embedded, I barely worked with embedded myself, hence I'm not sure about some of the embedded things). Or you could define:

type Query {
  findLetter(letter: String!): Letter
}

And then you could get the words via:

query GetWordsByLetter {
    findLetter(_size: 6, letter: "a"){
      data {
        words {
           ...
        }
      }
    }
 }

If that doesn't work for you, you probably want to specify a resolver on the query which will generate a User Defined Function for you. This requires you to dive into FQL since UDFs are specified in FQL but it will unlock much more possibilities.

(resolver docs: https://docs.fauna.com/fauna/current/api/graphql/directives/d_resolver)

Upvotes: 1

Related Questions