How to program an iterator in scala without using a mutable variable?

Question

I want to implement the iterator trait but in the functional way, ie, without using a var. How to do that?

Suppose I have an external library where I get some elements by calling a function getNextElements(numOfElements: Int):Array[String] and I want to implement an Iterator using that function but without using a variable indicating the "current" array (in my case, the var buffer). How can I implement that in the functional way?

class MyIterator[T](fillBuffer: Int => Array[T]) extends Iterator[T] {
    var buffer: List[T] = fillBuffer(10).toList

    override def hasNext(): Boolean = {
        if (buffer.isEmpty) buffer = fillBuffer(10).toList
        buffer.nonEmpty
    }

    override def next(): T = {
        if (!hasNext()) throw new NoSuchElementException()

        val elem: T = buffer.head
        buffer = buffer.tail
        elem
    }
}

class Main extends App {
    def getNextElements(num: Int): Array[String] = ???

    val iterator = new MyIterator[String](getNextElements)

    iterator.foreach(println)
}

Answer 1

Iterators are mutable, at least without an interface that also returns a state variable, so you can't in general implement the interface directly without some sort of mutation.

That being said, there are some very useful functions in the Iterator companion object that let you hide the mutation, and make the implementation cleaner. I would implement yours something like:

Iterator.continually(getNextElements(10)).flatten

This calls getNextElements(10) whenever it needs to fill the buffer. The flatten changes it from an Iterator[Array[A]] to an Iterator[A].

Note this returns an infinite iterator. Your question didn't say anything about detecting the end of your source elements, but I would usually implement that using takeWhile. For example, if getNextElements returns an empty array when there are no more elements, you can do:

 Iterator.continually(getNextElements(10)).takeWhile(!_.isEmpty).flatten