Continuing to work on Functional Programming in Scala exercises, I'm working on implementing foldRight on an IndexedSeq type.
Since foldRight will evaluate with right associativity, I created the following operator for pattern matching.
object :++ {
def unapply[T](s: Seq[T]) = s.lastOption.map(last =>
(last, s.take(s.length - 1)))
}
And then implemented as so:
object IndexedSeqFoldable extends Foldable[IndexedSeq] {
override def foldRight[A, B](as: IndexedSeq[A])(z: B)(f: (A, B) => B): B = {
def go(bs: Seq[A], acc: B): B = bs match {
case x :++ xs => go(xs, f(x, acc))
case _ => acc
}
go(as, z)
}
Ignoring the fact that foldRight can be written with foldLeft, how does my approach hold up?
Related
I'm learning Scala. Now I have this code snippet:
sealed abstract class BSTree {
def fold[A](init: A)(f: (A, Int) => A): A = this match {
case Empty => init
case Node(left, data, right) =>
val curInorder:A = f(left.fold(init)(f), data)
right.fold(curInorder)(f)
}
}
case object Empty extends BSTree
case class Node(left: BSTree, data: Int, right: BSTree) extends BSTree
My aim is to add another method toList in class BSTree, which is on top
of method fold and build a List from the binary tree's inorder traversal.
My current implementation is:
sealed abstract class BSTree {
def fold[A](init: A)(f: (A, Int) => A): = .....//code snippet skipped
def toList: List[Int] =
fold(Nil: List[Int])((xs: List[Int], hd)=> hd::xs).reverse
}
But I feel that building a List and then reversing it is ugly. Is there a more elegant approach?
Any hints are appreciated.
First of all, your fold is not tail recursive which for large input might result in StackOverflowException. I'd encourage you to try out and implement it on your own using Stack. For reference I'll place a sample implementation at the bottom of my post.
Secondly, as it was already mentioned in comments - you might want to use ListBuffer so that building your list is more efficient in reversed order (thus, there is no need to reverse it back).
Here's a one-liner:
def toList: List[Int] = fold(ListBuffer.empty[Int])(_ += _).toList
And the the reference for implementing tail-recursive fold:
def fold[B](init: B)(op: (B, A) => B): B = {
def go(stack: List[(A, Tree[A])], current: Tree[A], acc: B): B = (current, stack) match {
case (Empty, Nil) => acc
case (Empty, (d, r) :: xs) => go(xs, r, op(acc, d))
case (Node(l, d, r), _) => go((d, r) +: stack, l, acc)
}
go(Nil, this, init)
}
I find that simply using xs :+ hd instead of hd::xs puts the values in the correct order (depth-first, left-to-right).
val testTree: BSTree =
Node(Node(Empty, 0, Empty), 1, Node(Empty, 2, Node(Node(Empty, 3, Empty), 4, Empty)))
def toList(bSTree: BSTree): List[Int] =
bSTree.fold(List[Int]())((acc, next) => acc :+ next)
toList(testTree) // List(0,1,2,3,4)
My implementation above is O(n²). We can improve it to O(n) by using a ListBuffer, as per #dkim's comment, or we can use Vector and then convert to List when we're done.
Apart from simply fixing the toList method, we might ask why the result of using fold to implement toList didn't agree with our intuition (giving us a backwards list instead of a forwards list). One might point out that the fold signature for list matches the structure of the List class hierarchy.
abstract class List[+A] {
def fold[B](init: B)(step: (A, B) => B): B
}
case object Empty extends List[Nothing] {
def fold[B](init: B)(step: (A, B) => B): B = init
}
case class Cons[+A](head: A, tail: List[A]) extends List[A] {
def fold[B](init: B)(step: (A, B) => B): B =
step(head, tail.fold(init)(step))
}
Notice how the method signature of fold matches the class hierarchy, even down to the values that each implementing class holds. (Aside: For purposes of brevity, I am using a very naive implementation of fold that is neither efficient nor stack safe. Production implementations should be tail recursive or use a loop and a mutable buffer, but the point is that the method signature would be the same.)
We can do the same for your BSTree class, the fold signature would be:
abstract class BSTree {
def fold[A](withEmpty: A)(withNode: (A, Int, A) => A): A
}
Then toList would be tree.fold(List[Int]())((l, n, r) => l ++ List(n) ++ r). But again, use a buffer or Vector to get decent performance if you anticipate tree being even about 50 entries or so.
I'm implementing the List type in Scala when following a book.
Here's the definition of my List type:
sealed trait List[+A]
case object Nil extends List[Nothing]
case class Cons[+A](head: A, tail: List[A]) extends List[A]
All the later mentioned functions are defined in the companion object List in the same file
object List
I wrote foldLeft and foldRight as the following
def foldLeft[A,B](l: List[A], z: B)(f: (B, A) => B): B = l match {
case Nil => z
case Cons(x, xs) => foldLeft(xs, f(z, x))(f)
}
def foldRight[A,B](l: List[A], z: B)(f: (A, B) => B): B = l match {
case Nil => z
case Cons(x, xs) => f(x, foldRight(xs, z)(f))
}
There's an exercise on the book, which is to implement foldLeft using foldRight. Here's my initial implementation
def foldLeftWithRight[A,B](l: List[A], z: B)(f: (B, A) => B): B = {
foldRight(l, z)((a: A, b: B) => f(b, a))
}
Then I think I should write another function to do the reverse arguments if I'm to implement foldRight using foldLeft. As follows:
def reverseArgs[A,B](f: (A, B) => B): (B, A) => B = {
(b: B, a: A) => f(a, b)
}
So I changed code of foldLeftWithRight to the following:
def foldLeftWithRight[A,B](l: List[A], z: B)(f: (B, A) => B): B = {
foldRight(l, z)(reverseArgs(f))
}
And IntelliJ is complaining about reverseArgs(f):
Type mismatch: expected (A, B) => B, actual (B, B) => B
When I try to compile the code, the error is the following:
Error:(21, 37) type mismatch;
found : (B, A) => B
required: (B, Any) => Any
foldRight(l, z)(reverseArgs(f))
An interesting observation is that when I use the reverseArgs on foldRightWithLeft, there's no problem at all:
def foldRightWithLeft[A,B](l: List[A], z: B)(f: (A, B) => B): B = {
foldLeft(l, z)(reverseArgs(f))
}
What is going on here?
If you rename type parameters of your reverseArgs function to X and Y, you'll get something like
def reverseArgs[X ,Y](f: (X, Y) => Y): (Y, X) => Y = ???
Type of f in foldLeftWithRight is (B, A) => B. Passing that to reverseArgs means that:
X = B
Y = A
Y = B
I guess Intellij infers from here that A = B and this is why it's complaining that (B, B) => B isn't (A, B) => B. Scalac decides that Y = Any instead, because it's the least upper bound of two potentially unrelated types.
Good solution here is to generalize more. Return type of reversed function does not have to be one of parameter types, so you can introduce another generic type for that:
def reverseArgs[X ,Y, Z](f: (X, Y) => Z): (Y, X) => Z = {
(b: Y, a: X) => f(a, b)
}
In functional programmming, there are two important methods named foldLeft and foldRight. Here is the implementation of foldRight
sealed trait List[+A]
case object Nil extends List[Nothing]
case class Cons[+A](head: A, tails: List[A]) extends List[A]
def foldRight[A, B](ls: List[A], z: B)(f: (A, B) => B): B = ls match {
case Nil => z
case Cons(x, xs) => f(x, foldRight(xs, z)(f))
}
And here is the implementation of foldLeft:
#annotation.tailrec
def foldLeft[A, B](ls: List[A], z: B)(f: (B, A) => B): B = ls match {
case Nil => z
case Cons(x, xs) => foldLeft(xs, f(z, x))(f)
}
}
My question is: I read from many documents, they often put order of f function is: f: (B, A) => B instead of f: (A, B) => B. Why this definition is better? Because if we use otherwise way, it will have same signature with foldLeft, and it will be better.
Because foldLeft "turns around" the cons structure in its traversal:
foldRight(Cons(1, Cons(2, Nil)), z)(f) ~> f(1, f(2, z))
^ ^
A B
foldLeft(Cons(1, Cons(2, Nil)), z)(f) ~> f(f(z, 2), 1)
^ ^
B A
And since it visits the conses in opposite order, the types are also traditionally flipped. Of course the arguments could be swapped, but if you have a non-commutative operation, and expect the "traditional" behaviour, you'll be surprised.
...if we use otherwise way, it will have same signature with foldLeft, and it will be better.
No, I think that would not be better. If they had the same signature then the compiler would not be able to catch it if you intend to use one but accidentally type in the other.
The A/B order is also a handy reminder of where B (initial or "zero") value goes in relationship to the collection of A elements.
foldLeft: B-->>A, A, A, ... // (f: (B, A) => B)
foldRight: ... A, A, A<<--B // (f: (A, B) => B)
As part of an exercise from FP in Scala, I'm working on the implementation of foldLeft on an IndexedSeq.
I wrote 2 functions:
def foldLeft[A, B](as: IndexedSeq[A])(z: B)(f: (B, A) => B): B = {
def go(bs: IndexedSeq[A], acc: B): B = {
if (bs.isEmpty) acc
else go(bs.tail, f(acc, bs.head))
}
go(as, z)
}
And, then the pattern match way:
def foldLeftPM[A, B](as: IndexedSeq[A])(z: B)(f: (B, A) => B): B = {
def go(bs: IndexedSeq[A], acc: B): B = bs match {
case x +: xs => go(xs, f(acc, x))
case _ => acc
}
go(as, z)
}
EDIT Note that I got the +: operator from dhgs's answer. It appears to be a member of IndexedSeq's class or its parent since it's available without defining per the linked post.
Is either way better (from a performance or idiomatic Scala point of view)?
The pattern match is definitely more idiomatic.
For performance, they should be about the same, since they are exactly equivalent.
Though only benchmarking would decide, and that includes a lot of assumptions.
I feel insertSortRight is less efficient than insertSortLeft because insertSortRight needs to call List.last (which is O(n)) as one of the arguments to insert(), where insertSortLeft calls List.head (which is O(1)) as one of the arguments to insert().
Is this understanding correct? Thanks.
def insertSortRight(unsorted: List[Int]) : List[Int] = {
(unsorted :\ List[Int]()) ((a, b) => insert(a, b))
}
def insertSortLeft(unsorted: List[Int]) : List[Int] = {
(List[Int]() /: unsorted) ((a, b) => insert(b, a))
}
def insert(a: Int, list: List[Int]) : List[Int] = list match {
case List() => List(a)
case y::ys => if (a > y) y::insert(a, ys) else a::y::ys
}
DHG answered "always prefer left folding". But, Programming in Scala has an example the other way.
def flattenLeft[T](xss: List[List[T]]) = (List[T]() /: xss) (_ ::: )
def flattenRight[T](xss: List[List[T]]) = (xss :~List[T]()) ( ::: _)
I guess that is because flattenRight in this case is achieved by just one function call, while flattenLeft is achieved by n function call?
So, for a List, since head operations are desired, foldLeft is the natural choice. That way you work through the list from left to right, always taking the head. As you can see, its implementation (on LinearSeqOptimized) simply uses a while-loop and traverses once.
override /*TraversableLike*/
def foldLeft[B](z: B)(f: (B, A) => B): B = {
var acc = z
var these = this
while (!these.isEmpty) {
acc = f(acc, these.head)
these = these.tail
}
acc
}
It seems like 'foldRight' would be O(n^2) since, in order to take the last element, you have to traverse the n elements of the List n times, but the library actually optimizes this for you. Behind the scenes, foldRight is implemented like this (also on LinearSeqOptimized):
def foldRight[B](z: B)(f: (A, B) => B): B =
if (this.isEmpty) z
else f(head, tail.foldRight(z)(f))
As you can see, this function is constructed by recursively calling foldRight on the tail, holding each head on the stack, and applying the function to each head in reverse order after reaching the last element.