Reputation: 13918
Consider this list composed of objects which are instances of case classes:
A, B, Opt(A),C, Opt(D), F, Opt(C), G, Opt(H)
I wan to normalize this list to get this result:
A, B, C, Opt(D), F, G, Opt(H)
As you see, if there are elements A
and Opt(A)
I replace them with just A
or said other way, I have to remove OPT(A)
element.
I would like:
Upvotes: 0
Views: 142
Reputation: 24423
This might be a little more concise, as filtering is what you want ;-):
scala> List(1,2,3,Some(4),5,Some(5))
res0: List[Any] = List(1, 2, 3, Some(4), 5, Some(5))
scala> res0.filter {
| case Some(x) => !res0.contains(x)
| case _ => true
| }
res1: List[Any] = List(1, 2, 3, Some(4), 5)
edit: For large collections it might be good to use a toSet
or directly use a Set
.
Upvotes: 2
Reputation: 51109
If you don't care about order then efficiency leads you to use a Set:
xs.foldLeft(Set.empty[Any])({ case (set, x) => x match {
case Some(y) => if (set contains y) set else set + x
case y => if (set contains Some(y)) set - Some(y) + y else set + y
}}).toList
Alternatively:
val (opts, ints) = xs.toSet.partition(_.isInstanceOf[Option[_]])
opts -- (ints map (Option(_))) ++ ints toList
Upvotes: 1
Reputation: 92056
Not the most efficient solution, but certainly a simple one.
scala> case class Opt[A](a: A)
defined class Opt
scala> val xs = List(1, 2, Opt(1), 3, Opt(4), 6, Opt(3), 7, Opt(8))
xs: List[Any] = List(1, 2, Opt(1), 3, Opt(4), 6, Opt(3), 7, Opt(8))
scala> xs flatMap {
| case o @ Opt(x) => if(xs contains x) None else Some(o)
| case x => Some(x)
| }
res5: List[Any] = List(1, 2, 3, Opt(4), 6, 7, Opt(8))
Upvotes: 2