Querying secondary relationships slow as db size increases

[david-on-github]

Responses to queries with secondary relationships become incrementally slower relative to the size of the db. (Including unrelated collections)

Example Schema

type Dog {
  name: String
} 

type Director {
  name: String 
  movies: [Movie] @relation(name: "directorMovies")
}

type Movie {
  title: String
  director: Director @primary @relation(name: "directorMovies")
}

*Note: Dogs is not related to Director or Movies + Not included in this query:

query {
  Director {
   movies {
     _docID
   } 
  }
}

Calling the above query gives the following response times (where number of documents in the db is):

Db seeded with:  0 Directors with 0 movies

0 Dogs 22ms
1000 Dogs 23ms
2000 Dogs 31ms
3000 Dogs 37ms
4000 Dogs 30ms
5000 Dogs 46ms

Db seeded with: 100 Directors with 1 movie each

0 Dogs 124ms
1000 Dogs 307ms
2000 Dogs 625ms
3000 Dogs 727ms
4000 Dogs 921ms
5000 Dogs 1177ms

Db seeded with: 200 Directors with 0 movies

0 Dogs 116ms
1000 Dogs 513ms
2000 Dogs 884ms
3000 Dogs 1322ms
4000 Dogs 1788ms
5000 Dogs 2199ms