Scala模式：对于产生未来的理解力[A]

Question

What is the pattern used in Scala to deal with the scenario:

You have a bunch of futures (they can be whatever, but for the sake of example...)

val aF = Future { true }
val bF = Future { Option(3) }
val cF = Future { myObject }

and you have some function that returns a future

def fooF: Future[SomeObject]

I want to do something like:

for {
    a <- aF
    b <- bF
    c <- cF
} yield {
    if (a) {
        // do stuff with b & c
        fooF
    } else {
        Future.successful(SomeObject)
    }
}

I want to return a value of Future[SomeObject] , but I call fooF inside of the yield statement, I will get a Future[Future[SomeObject]]

Answer 1

Here is another solution :

def doStuffWith(a: A, b: B, c: C): Future[SomeObject] = if (a) {
  // do stuff with b & c
  fooF
} else Future.successful(SomeObject)

for {
  a <- aF
  b <- bF
  c <- cF
  d <- doStuffWith(a, b, c)
} yield d

As discussed in @laughedelic answer, this is a subjective view, but I believe this way to be more readable, and maintainable (taking out the function always to unit test it, for instance).

Answer 2

• In Scala 2.12 Future has a flatten method:

  Creates a new future with one level of nesting flattened, this method is equivalent to flatMap(identity).

  So you can write for { ... } yield { ... } flatten .

• In Scala <2.12 you can achieve the same with flatMap(identity) (as mentioned above)

---

An alternative solution is to use flatMap instead of for-comprehension:

aF.flatMap { a =>
  if (a) {
    (bF zip cF).flatMap { (b, c) =>
      // do stuff with b & c
      fooF
    }
  } else 
    Future.successful(SomeObject)
  }
}