Question about @recurse with expand(_all_)

Users Dgraph Cloud
, ,
1

I hope this is the right category for this question.

I was recently experimenting with the @recurse directive in combination with expand(_all_). I found that doing so very quickly gets you in big troubles since (at least for the Shared Cluster Instance), the chance of getting a timeout error is pretty high. This obviously depends on the amount of leaves a the root node has you want to query.

However, I found that

{
  expRec(func: uid(0x1)) @recurse (loop: false, depth: 2) {
    uid
    expand(_all_)
  }
}

leads to substantially more timeouts than doing

{
  expAll(func: uid(0x1)) {
    uid
   expand(_all_) {
     uid
     expand(_all_)
   }
  }
}

which, at least in my opinion, should produce exactly the same results and take more or less the same amount of time. Since I cannot really give any examples here (my types are way to big an to nested to put them in here), I know it might be tricky to reproduce.

Motivation

The motivation behind this relies in the mutation response from a custom lambda, in combination with Relay. E.g. if I want to update a type User, the return type has to be of type User as well. If, like in my case, you have a huge object, it becomes pretty hard to maintain which fields should be returned by the lambda and which not. So I have tried to simply return the most fields possible for a certain type in order not to run in any Relay errors.

Is there anyone else having the same issues?