A TC39 proposal to add a method to iterators for producing an iterator of its subsequences.

Stage: 0

See the January 2024 presentation to committee.

It can be useful to consume a stream by more than one value at a time. For example, certain algorithms require looking at adjacent elements.

A common solution for this is a chunks method that works like the following:

let digits = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9].values();

let chunksOf2 = Array.from(digits.chunks(2));
// [ [0, 1], [2, 3], [4, 5], [6, 7], [8, 9] ]

let chunksOf3 = Array.from(digits.chunks(3));
// [ [0, 1, 2], [3, 4, 5], [6, 7, 8], [9] ]

let chunksOf4 = Array.from(digits.chunks(4));
// [ [0, 1, 2, 3], [4, 5, 6, 7], [8, 9] ]

A more flexible solution is a sliding window method, usually named windows:

let windowsOf3 = Array.from(digits.windows(3));
// [ [0, 1, 2], [1, 2, 3], [2, 3, 4], [3, 4, 5], [4, 5, 6], [5, 6, 7], [6, 7, 8], [7, 8, 9] ]

let windowsOf2AdvancingBy3 = Array.from(digits.windows(2, 3));
// [ [0, 1], [3, 4], [6, 7], [9] ]

chunks is just a specialisation of windows where chunks(n) is equivalent to windows(n, n).

let chunksOf4 = Array.from(digits.windows(4, 4));
// [ [0, 1, 2, 3], [4, 5, 6, 7], [8, 9] ]

• sliding windows sliding in/out past ends instead of sliding within boundaries?
• default chunk size to 2?
• chunking by 0 length?
• windows advancing by 0?
• should truncated chunks be returned?
  • filler/padding elements?
• should windows ever be truncated?
  • should preserve equivalence with chunks, so answer should match chunks
  • only when full iterator doesn't fill a single window?
  • just lock step to 1 for windowing?