Idiomatic Scala way to traverse Iterator-like data structure

Question

Given a class from a legacy library that behaves like an Iterator such that you can traverse it's content with hasNext and next but it does not implement the Iterator interface:

class LegacyIterator[T](iterable: Iterable[T]) {
  val iterator: Iterator[T] = iterable.iterator
  def hasNext: Boolean = iterator.hasNext
  def next(): T = iterator.next()
}

val lagIter: LegacyIterator[Int] = new LegacyIterator(List(1, 2, 3))

What would be your opinion on the most elegant way to traverse such a data structure - idially inside the scope of the function? I've come up with two versions:

One using @tailrec

@tailrec
def lagIterList(accu: List[Int]): List[Int] =
  if (lagIter.hasNext) lagIterList(lagIter.next() :: accu) else accu

val res = lagIterList(Nil).reverse

And one using LazyList

def ll: LazyList[Int] = if(lagIter.hasNext) lagIter.next() #:: ll else LazyList.empty
val res = ll.toList

What would you think would be more Scala idiomatic? The tailrec version is probably slightly faster on larger data structures but the LazyList version makes better use of the collections library.

Or any other ideas?

Lesiak · Accepted Answer · 2023-02-28 11:23:02Z

2

I would consider an implicit wrapper class extending Iterator[T]:

implicit class LegacyIteratorWrapper[T](private val underlying: LegacyIterator[T]) extends Iterator[T] {
  override def hasNext: Boolean = underlying.hasNext
  override def next(): T = underlying.next()
}

val lagIter: LegacyIterator[Int] = new LegacyIterator(List(1, 2, 3))

// traversal
for (elem <- lagIter) println(elem)

// conversion to list
val lagIterList = lagIter.toList

answered Feb 28, 2023 at 11:23

Lesiak

26.5k2 gold badges54 silver badges79 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Zwackelmann Over a year ago

Very nice solution. Since I only have one spot where I have to apply this conversion, I was more looking for a slick, nifty solution to be applied inside my functions scope. In general however, I would agree that wrapping the LegacyIterator with an implicit class to gain full access to the collection API, is quite elegant.

Lesiak Over a year ago

What wasn't clear to me from your question is whether you can use the LegacyIterator api directly to traverse it, or if you need to pass it to some other method expecting Iterator.

Zwackelmann Over a year ago

True, that aspect was not clear. I'll try to make my case more clear and update my question shortly. I was mainly interesed in opinions about weather people would prefer the tailrec or LazyList version to traverse given iterator inside a functions scope. Or if both versions look ugly from a purely subjective point of view.

Lesiak Over a year ago

This might be old-school, but I would turn to a while loop if you only care about traversal in one function

Dima · Accepted Answer · 2023-02-28 14:43:50Z

2

Just convert it into a normal iterator:

val normalIterator = Iterator.continually(legacyIterator)
  .takeWhile(_.hasNext)
  .map(_.next)

Then you can use any of the regular scala tooling to traverse/transform it, e.g.: val list = normalIterator.toList

answered Feb 28, 2023 at 14:43

Dima

40.6k6 gold badges54 silver badges80 bronze badges

1 Comment

Zwackelmann Over a year ago

Cool! Thanks for the input. Looks very clean to me.

Luis Miguel Mejía Suárez · Accepted Answer · 2023-02-28 15:38:32Z

1

Well, it depends on what you mean by "traverse it". If you only mean, creating a List out of it, I would just use unfold

def toList[T](iter: LegacyIterator[T]): List[T] =
  List.unfold(()) { _ =>
    Option.when(iter.hasNext()) {
      iter.next() -> ()
    }
  }

edited Feb 28, 2023 at 15:38

answered Feb 28, 2023 at 15:04

Luis Miguel Mejía Suárez

23k4 gold badges31 silver badges50 bronze badges

1 Comment

Zwackelmann Over a year ago

Cool! Thanks! I didn't know unfold before. I saw it was introduced in 2.13. Very nifty solution.

Collectives™ on Stack Overflow

Idiomatic Scala way to traverse Iterator-like data structure

3 Answers 3

4 Comments

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related