Array.fold

Reduces the elements of this array into a single value by calling a function on each element, starting from the left.

Signature

  fun fold[A, B](
╰╴No statement matched
      self[A],
      B init,
      (B acc, A next) -> (B) transform_function,
  ) -> B
thp

Parameters

• self: The callee.
• init: The initial value of the fold.
• transform_function: The function to apply to every element.

Return value

• If self is an empty array, returns init
• Otherwise, returns the result of applying transform_function to the accumulator acc, and the next element next.

Description

Fold allows you to transform an array of A into a single B, following an arbitraty function.

For example, let’s say that you have an array of numbers:

val digits = [1, 9, 9, 5]
╰╴No statement matched
thp

And you want to join all digits into a String like "1985". You can achieve this with a fold.

val digits = [1, 9, 8, 5]
╰╴No statement matched
val init = ""
fun f(String acc, Int next) = acc ++ next

digits.fold("", f) // "1985"
thp

• f is a function that takes a previous value (known as the accumulator) and a new value (known as next) and concatenates them. So f("", 1) results in "1", f("1", 2) in "12" and so on.

• To begin, the accumulator is set to init, in this case ""
• First, f is called with the accumulator ("") and the first element of the array (1) resulting in f("", 1)
• This results in the string "1". This becomes the new value of the accumulator.
• Then f is called with the accumulator and the 2nd element of the array. This results in f("1", 9), which is "19". This is now the accumulator.
• Then, f is called the same way as before, resulting in f("19", 8). The accumulator now is "198".
• Finally, f is called the same way, resulting in f("198", 5) which is "1985". This is the value that fold returns.

A sort of visualization is the following:

[1, 9, 8, 5].fold("", f) =
    f( f( f( f("", 1), 9), 8), 5)
    f( f( f(    "1"  , 9), 8), 5)
    f( f(          "19"  , 8), 5)
    f(                "198"  , 5)
                         "1985"

Hopefully now you understand how a fold works, and we can continue using the abstraction.

The transform_function can be any function, and can operate over any type.

For example, you could sum all numbers instead of concatenating like this:

val digits = [1, 9, 8, 5]
╰╴No statement matched
digits.fold(0, fun(acc, next) = acc + next)	// 23
thp

f( f( f( f(0, 1), 9), 8), 5)
f( f( f(    1   , 9), 8), 5)
f( f(          10   , 8), 5)
f(                 18   , 5)
                       23

Or you could pass the digits to a builder class:

val digits = [1, 9, 8, 5]
╰╴No statement matched
val builder = ThingBuilder()

digits.fold(builder, fun(acc, next) = acc.addNumber(next))

val result = builder.build()
thp

Or anything.