gleam/iterator

Types

An iterator is a lazily evaluated sequence of element.

Iterators are useful when working with collections that are too large to fit in memory (or those that are infinite in size) as they only require the elements currently being processed to be in memory.

As a lazy data structure no work is done when an iterator is filtered, mapped, etc, instead a new iterator is returned with these transformations applied to the stream. Once the stream has all the required transformations applied it can be evaluated using functions such as fold and to_list.

pub opaque type Iterator(element)
pub type Step(element, accumulator) {
  Next(element: element, accumulator: accumulator)
  Done
}

Constructors

  • Next(element: element, accumulator: accumulator)
  • Done

Functions

pub fn all(
  in iterator: Iterator(a),
  satisfying predicate: fn(a) -> Bool,
) -> Bool

Returns True if all elements emitted by the iterator satisfy the given predicate, False otherwise.

This function short-circuits once it finds a non-satisfying element.

An empty iterator results in True.

Examples

empty()
|> all(fn(n) { n % 2 == 0 })
// -> True
from_list([2, 4, 6, 8])
|> all(fn(n) { n % 2 == 0 })
// -> True
from_list([2, 4, 5, 8])
|> all(fn(n) { n % 2 == 0 })
// -> False
pub fn any(
  in iterator: Iterator(a),
  satisfying predicate: fn(a) -> Bool,
) -> Bool

Returns True if any element emitted by the iterator satisfies the given predicate, False otherwise.

This function short-circuits once it finds a satisfying element.

An empty iterator results in False.

Examples

empty()
|> any(fn(n) { n % 2 == 0 })
// -> False
from_list([1, 2, 5, 7, 9])
|> any(fn(n) { n % 2 == 0 })
// -> True
from_list([1, 3, 5, 7, 9])
|> any(fn(n) { n % 2 == 0 })
// -> False
pub fn append(
  to first: Iterator(a),
  suffix second: Iterator(a),
) -> Iterator(a)

Appends two iterators, producing a new iterator.

This function does not evaluate the elements of the iterators, the computation is performed when the resulting iterator is later run.

Examples

from_list([1, 2])
|> append(from_list([3, 4]))
|> to_list
// -> [1, 2, 3, 4]
pub fn at(
  in iterator: Iterator(a),
  get index: Int,
) -> Result(a, Nil)

Returns nth element yielded by the given iterator, where 0 means the first element.

If there are not enough elements in the iterator, Error(Nil) is returned.

For any index less than 0 this function behaves as if it was set to 0.

Examples

from_list([1, 2, 3, 4]) |> at(2)
// -> Ok(3)
from_list([1, 2, 3, 4]) |> at(4)
// -> Error(Nil)
empty() |> at(0)
// -> Error(Nil)
pub fn chunk(
  over iterator: Iterator(a),
  by f: fn(a) -> b,
) -> Iterator(List(a))

Creates an iterator that emits chunks of elements for which f returns the same value.

Examples

from_list([1, 2, 2, 3, 4, 4, 6, 7, 7])
|> chunk(by: fn(n) { n % 2 })
|> to_list
// -> [[1], [2, 2], [3], [4, 4, 6], [7, 7]]
pub fn concat(iterators: List(Iterator(a))) -> Iterator(a)

Joins a list of iterators into a single iterator.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

[[1, 2], [3, 4]]
|> map(from_list)
|> concat
|> to_list
// -> [1, 2, 3, 4]
pub fn cycle(iterator: Iterator(a)) -> Iterator(a)

Creates an iterator that repeats a given iterator infinitely.

Examples

from_list([1, 2])
|> cycle
|> take(6)
|> to_list
// -> [1, 2, 1, 2, 1, 2]
pub fn drop(
  from iterator: Iterator(a),
  up_to desired: Int,
) -> Iterator(a)

Evaluates and discards the first N elements in an iterator, returning a new iterator.

If the iterator does not have enough elements an empty iterator is returned.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

from_list([1, 2, 3, 4, 5])
|> drop(up_to: 3)
|> to_list
// -> [4, 5]
from_list([1, 2])
|> drop(up_to: 3)
|> to_list
// -> []
pub fn drop_while(
  in iterator: Iterator(a),
  satisfying predicate: fn(a) -> Bool,
) -> Iterator(a)

Creates an iterator that drops elements while the predicate returns True, and then yields the remaining elements.

Examples

from_list([1, 2, 3, 4, 2, 5])
|> drop_while(satisfying: fn(x) { x < 4 })
|> to_list
// -> [4, 2, 5]
pub fn each(
  over iterator: Iterator(a),
  with f: fn(a) -> b,
) -> Nil

Traverse an iterator, calling a function on each element.

Examples

empty() |> each(io.println)
// -> Nil
from_list(["Tom", "Malory", "Louis"]) |> each(io.println)
// -> Nil
// Tom
// Malory
// Louis
pub fn empty() -> Iterator(a)

Creates an iterator that yields no elements.

Examples

empty() |> to_list
// -> []
pub fn filter(
  iterator: Iterator(a),
  keeping predicate: fn(a) -> Bool,
) -> Iterator(a)

Creates an iterator from an existing iterator and a predicate function.

The new iterator will contain elements from the first iterator for which the given function returns True.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

import gleam/int
from_list([1, 2, 3, 4])
|> filter(int.is_even)
|> to_list
// -> [2, 4]
pub fn find(
  in haystack: Iterator(a),
  one_that is_desired: fn(a) -> Bool,
) -> Result(a, Nil)

Finds the first element in a given iterator for which the given function returns True.

Returns Error(Nil) if the function does not return True for any of the elements.

Examples

find(from_list([1, 2, 3]), fn(x) { x > 2 })
// -> Ok(3)
find(from_list([1, 2, 3]), fn(x) { x > 4 })
// -> Error(Nil)
find(empty(), fn(_) { True })
// -> Error(Nil)
pub fn first(from iterator: Iterator(a)) -> Result(a, Nil)

Returns the first element yielded by the given iterator, if it exists, or Error(Nil) otherwise.

Examples

from_list([1, 2, 3]) |> first
// -> Ok(1)
empty() |> first
// -> Error(Nil)
pub fn flat_map(
  over iterator: Iterator(a),
  with f: fn(a) -> Iterator(b),
) -> Iterator(b)

Creates an iterator from an existing iterator and a transformation function.

Each element in the new iterator will be the result of calling the given function on the elements in the given iterator and then flattening the results.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

from_list([1, 2])
|> flat_map(fn(x) { from_list([x, x + 1]) })
|> to_list
// -> [1, 2, 2, 3]
pub fn flatten(iterator: Iterator(Iterator(a))) -> Iterator(a)

Flattens an iterator of iterators, creating a new iterator.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

from_list([[1, 2], [3, 4]])
|> map(from_list)
|> flatten
|> to_list
// -> [1, 2, 3, 4]
pub fn fold(
  over iterator: Iterator(a),
  from initial: b,
  with f: fn(b, a) -> b,
) -> b

Reduces an iterator of elements into a single value by calling a given function on each element in turn.

If called on an iterator of infinite length then this function will never return.

If you do not care about the end value and only wish to evaluate the iterator for side effects consider using the run function instead.

Examples

from_list([1, 2, 3, 4])
|> fold(from: 0, with: fn(acc, element) { element + acc })
// -> 10
pub fn fold_until(
  over iterator: Iterator(a),
  from initial: b,
  with f: fn(b, a) -> ContinueOrStop(b),
) -> b

Like fold, fold_until reduces an iterator of elements into a single value by calling a given function on each element in turn, but uses list.ContinueOrStop to determine whether or not to keep iterating.

If called on an iterator of infinite length then this function will only ever return if the function returns list.Stop.

Examples

import gleam/list

let f = fn(acc, e) {
  case e {
    _ if e < 4 -> list.Continue(e + acc)
    _ -> list.Stop(acc)
  }
}

from_list([1, 2, 3, 4])
|> fold_until(from: acc, with: f)
// -> 6
pub fn from_list(list: List(a)) -> Iterator(a)

Creates an iterator that yields each element from the given list.

Examples

from_list([1, 2, 3, 4])
|> to_list
// -> [1, 2, 3, 4]
pub fn group(
  in iterator: Iterator(a),
  by key: fn(a) -> b,
) -> Dict(b, List(a))

Returns a Dict(k, List(element)) of elements from the given iterator grouped with the given key function.

The order within each group is preserved from the iterator.

Examples

from_list([1, 2, 3, 4, 5, 6])
|> group(by: fn(n) { n % 3 })
// -> dict.from_list([#(0, [3, 6]), #(1, [1, 4]), #(2, [2, 5])])
pub fn index(over iterator: Iterator(a)) -> Iterator(#(a, Int))

Wraps values yielded from an iterator with indices, starting from 0.

Examples

from_list(["a", "b", "c"]) |> index |> to_list
// -> [#("a", 0), #("b", 1), #("c", 2)]
pub fn interleave(
  left: Iterator(a),
  with right: Iterator(a),
) -> Iterator(a)

Creates an iterator that alternates between the two given iterators until both have run out.

Examples

from_list([1, 2, 3, 4])
|> interleave(from_list([11, 12, 13, 14]))
|> to_list
// -> [1, 11, 2, 12, 3, 13, 4, 14]
from_list([1, 2, 3, 4])
|> interleave(from_list([100]))
|> to_list
// -> [1, 100, 2, 3, 4]
pub fn intersperse(
  over iterator: Iterator(a),
  with elem: a,
) -> Iterator(a)

Creates an iterator that yields the given elem element between elements emitted by the underlying iterator.

Examples

empty()
|> intersperse(with: 0)
|> to_list
// -> []
from_list([1])
|> intersperse(with: 0)
|> to_list
// -> [1]
from_list([1, 2, 3, 4, 5])
|> intersperse(with: 0)
|> to_list
// -> [1, 0, 2, 0, 3, 0, 4, 0, 5]
pub fn iterate(
  from initial: a,
  with f: fn(a) -> a,
) -> Iterator(a)

Creates an iterator that inifinitely applies a function to a value.

Examples

iterate(1, fn(n) { n * 3 }) |> take(5) |> to_list
// -> [1, 3, 9, 27, 81]
pub fn last(iterator: Iterator(a)) -> Result(a, Nil)

Returns the last element in the given iterator.

Returns Error(Nil) if the iterator is empty.

This function runs in linear time.

Examples

empty() |> last
// -> Error(Nil)
range(1, 10) |> last
// -> Ok(9)
pub fn length(over iterator: Iterator(a)) -> Int

Counts the number of elements in the given iterator.

This function has to traverse the entire iterator to count its elements, so it runs in linear time.

Examples

empty() |> length
// -> 0
from_list([1, 2, 3, 4]) |> length
// -> 4
pub fn map(
  over iterator: Iterator(a),
  with f: fn(a) -> b,
) -> Iterator(b)

Creates an iterator from an existing iterator and a transformation function.

Each element in the new iterator will be the result of calling the given function on the elements in the given iterator.

This function does not evaluate the elements of the iterator, the computation is performed when the iterator is later run.

Examples

from_list([1, 2, 3])
|> map(fn(x) { x * 2 })
|> to_list
// -> [2, 4, 6]
pub fn map2(
  iterator1: Iterator(a),
  iterator2: Iterator(b),
  with fun: fn(a, b) -> c,
) -> Iterator(c)

Combines two interators into a single one using the given function.

If an iterator is longer than the other the extra elements are dropped.

This function does not evaluate the elements of the two iterators, the computation is performed when the resulting iterator is later run.

Examples

let first = from_list([1, 2, 3])
let second = from_list([4, 5, 6])
map2(first, second, fn(x, y) { x + y }) |> to_list
// -> [5, 7, 9]
let first = from_list([1, 2])
let second = from_list(["a", "b", "c"])
map2(first, second, fn(i, x) { #(i, x) }) |> to_list
// -> [#(1, "a"), #(2, "b")]
pub fn once(f: fn() -> a) -> Iterator(a)

Creates an iterator that yields exactly one element provided by calling the given function.

Examples

once(fn() { 1 }) |> to_list
// -> [1]
pub fn range(from start: Int, to stop: Int) -> Iterator(Int)

Creates an iterator of ints, starting at a given start int and stepping by one to a given end int.

Examples

range(from: 1, to: 5) |> to_list
// -> [1, 2, 3, 4, 5]
range(from: 1, to: -2) |> to_list
// -> [1, 0, -1, -2]
range(from: 0, to: 0) |> to_list
// -> [0]
pub fn reduce(
  over iterator: Iterator(a),
  with f: fn(a, a) -> a,
) -> Result(a, Nil)

This function acts similar to fold, but does not take an initial state. Instead, it starts from the first yielded element and combines it with each subsequent element in turn using the given function. The function is called as f(accumulator, current_element).

Returns Ok to indicate a successful run, and Error if called on an empty iterator.

Examples

from_list([])
|> reduce(fn(acc, x) { acc + x })
// -> Error(Nil)
from_list([1, 2, 3, 4, 5])
|> reduce(fn(acc, x) { acc + x })
// -> Ok(15)
pub fn repeat(x: a) -> Iterator(a)

Creates an iterator that returns the same value infinitely.

Examples

repeat(10)
|> take(4)
|> to_list
// -> [10, 10, 10, 10]
pub fn repeatedly(f: fn() -> a) -> Iterator(a)

Creates an iterator that yields values created by calling a given function repeatedly.

pub fn run(iterator: Iterator(a)) -> Nil

Evaluates all elements emitted by the given iterator. This function is useful for when you wish to trigger any side effects that would occur when evaluating the iterator.

pub fn scan(
  over iterator: Iterator(a),
  from initial: b,
  with f: fn(b, a) -> b,
) -> Iterator(b)

Creates an iterator from an existing iterator and a stateful function.

Specifically, this behaves like fold, but yields intermediate results.

Examples

// Generate a sequence of partial sums
from_list([1, 2, 3, 4, 5])
|> scan(from: 0, with: fn(acc, el) { acc + el })
|> to_list
// -> [1, 3, 6, 10, 15]
pub fn single(elem: a) -> Iterator(a)

Creates an iterator that yields the given element exactly once.

Examples

single(1) |> to_list
// -> [1]
pub fn sized_chunk(
  over iterator: Iterator(a),
  into count: Int,
) -> Iterator(List(a))

Creates an iterator that emits chunks of given size.

If the last chunk does not have count elements, it is yielded as a partial chunk, with less than count elements.

For any count less than 1 this function behaves as if it was set to 1.

Examples

from_list([1, 2, 3, 4, 5, 6])
|> sized_chunk(into: 2)
|> to_list
// -> [[1, 2], [3, 4], [5, 6]]
from_list([1, 2, 3, 4, 5, 6, 7, 8])
|> sized_chunk(into: 3)
|> to_list
// -> [[1, 2, 3], [4, 5, 6], [7, 8]]
pub fn step(iterator: Iterator(a)) -> Step(a, Iterator(a))

Eagerly accesses the first value of an iterator, returning a Next that contains the first value and the rest of the iterator.

If called on an empty iterator, Done is returned.

Examples

let assert Next(first, rest) = from_list([1, 2, 3, 4]) |> step

first
// -> 1

rest |> to_list
// -> [2, 3, 4]
empty() |> step
// -> Done
pub fn take(
  from iterator: Iterator(a),
  up_to desired: Int,
) -> Iterator(a)

Creates an iterator that only yields the first desired elements.

If the iterator does not have enough elements all of them are yielded.

Examples

from_list([1, 2, 3, 4, 5])
|> take(up_to: 3)
|> to_list
// -> [1, 2, 3]
from_list([1, 2])
|> take(up_to: 3)
|> to_list
// -> [1, 2]
pub fn take_while(
  in iterator: Iterator(a),
  satisfying predicate: fn(a) -> Bool,
) -> Iterator(a)

Creates an iterator that yields elements while the predicate returns True.

Examples

from_list([1, 2, 3, 2, 4])
|> take_while(satisfying: fn(x) { x < 3 })
|> to_list
// -> [1, 2]
pub fn to_list(iterator: Iterator(a)) -> List(a)

Evaluates an iterator and returns all the elements as a list.

If called on an iterator of infinite length then this function will never return.

Examples

from_list([1, 2, 3])
|> map(fn(x) { x * 2 })
|> to_list
// -> [2, 4, 6]
pub fn transform(
  over iterator: Iterator(a),
  from initial: b,
  with f: fn(b, a) -> Step(c, b),
) -> Iterator(c)

Creates an iterator from an existing iterator and a stateful function that may short-circuit.

f takes arguments acc for current state and el for current element from underlying iterator, and returns either Next with yielded element and new state value, or Done to halt the iterator.

Examples

Approximate implementation of index in terms of transform:

from_list(["a", "b", "c"])
|> transform(0, fn(i, el) { Next(#(i, el), i + 1) })
|> to_list
// -> [#(0, "a"), #(1, "b"), #(2, "c")]
pub fn try_fold(
  over iterator: Iterator(a),
  from initial: b,
  with f: fn(b, a) -> Result(b, c),
) -> Result(b, c)

A variant of fold that might fail.

The folding function should return Result(accumulator, error). If the returned value is Ok(accumulator) try_fold will try the next value in the iterator. If the returned value is Error(error) try_fold will stop and return that error.

Examples

from_list([1, 2, 3, 4])
|> try_fold(0, fn(acc, i) {
  case i < 3 {
    True -> Ok(acc + i)
    False -> Error(Nil)
  }
})
// -> Error(Nil)
pub fn unfold(
  from initial: a,
  with f: fn(a) -> Step(b, a),
) -> Iterator(b)

Creates an iterator from a given function and accumulator.

The function is called on the accumulator and returns either Done, indicating the iterator has no more elements, or Next which contains a new element and accumulator. The element is yielded by the iterator and the new accumulator is used with the function to compute the next element in the sequence.

Examples

unfold(from: 5, with: fn(n) {
 case n {
   0 -> Done
   n -> Next(element: n, accumulator: n - 1)
 }
})
|> to_list
// -> [5, 4, 3, 2, 1]
pub fn yield(
  element: a,
  next: fn() -> Iterator(a),
) -> Iterator(a)

Add a new element to the start of an iterator.

This function is for use with use expressions, to replicate the behaviour of the yield keyword found in other languages.

Examples

let iterator = {
  use <- yield(1)
  use <- yield(2)
  use <- yield(3)
  empty()
}
iterator |> to_list
// -> [1, 2, 3]
pub fn zip(
  left: Iterator(a),
  right: Iterator(b),
) -> Iterator(#(a, b))

Zips two iterators together, emitting values from both until the shorter one runs out.

Examples

from_list(["a", "b", "c"])
|> zip(range(20, 30))
|> to_list
// -> [#("a", 20), #("b", 21), #("c", 22)]
Search Document