ilist: Optimised list functions for doing index-related things

[ bsd3, library, list, mpl ] [ Propose Tags ]

Optimised list functions for doing index-related things. They're faster than common idioms in all cases, they avoid space leaks, and sometimes they fuse better as well.

[Skip to Readme]
Versions [faq],,,,
Change log
Dependencies base (>=4.10 && <4.14) [details]
License MPL-2.0
Copyright 2016-2019 Artyom Kazak (BSD-3-Clause) 2019-2020 Kowainik (MPL-2.0)
Author Artyom
Maintainer Kowainik <>
Category List
Home page
Bug tracker
Source repo head: git clone
Uploaded by shersh at Thu Dec 26 19:49:43 UTC 2019
Distributions LTSHaskell:, NixOS:, Stackage:
Downloads 3625 total (300 in the last 30 days)
Rating (no votes yet) [estimated by Bayesian average]
Your Rating
  • λ
  • λ
  • λ
Status Hackage Matrix CI
Docs available [build log]
Last success reported on 2019-12-26 [all 1 reports]


[Index] [Quick Jump]


Maintainer's Corner

For package maintainers and hackage trustees

Readme for ilist-

[back to package description]


Travis AppVeyor Hackage Stackage LTS Stackage Nightly MPL-2.0 license

What is this

This is a library with lots of list functions that are related to indices. It has often-reinvented deleteAt, setAt, etc, as well as indexed variants of functions from Data.List (e.g. imap, ifilter, izipWith). It has no dependencies, builds in about a second, and works on GHC from 7.4 to 8.0; the functions are optimised and benchmarked (for instance, the zip [0..] idiom is usually twice as slow, and sometimes 20× as slow).

So, this library is intended to be the canonical place for index-related functions. You are encouraged to depend on this library instead of reinventing the functions, using zip [0..], or using lens when all you need is a simple imap or ifoldr (not to mention that lens variants are usually 2–10 times slower for lists).

Why should you care

You shouldn't, actually. This is a small library, it won't change anyone's life, and if you care about speed you probably shouldn't be using lists anyway (unless you keep your fingers crossed and hope that fusion will kick in). So, consider it more of a public service announcement – “hey, just in case you ever need them, index-related functions live here”.


Unfortunately, Data.List.Indexed was taken by IndexedList, which implements such exciting things as “counted lists” and “conic lists”. Nope, I'm not bitter at all. Okay, maybe a bit, even tho it's completely unfair to IndexedList. Anyway:

import Data.List.Index

And you can use functions from Data.List by prepending i to them. There's also indexed :: [a] -> [(Int,a)] and a family of functions for modifying the element at an index (deleteAt, setAt, modifyAt, updateAt, insertAt).

Watch out – ifoldl has the index as the second parameter of the function:

ifoldl :: (b -> Int -> a -> b) -> b -> [a] -> b

That's the same convention that containers and vector use. Other functions pass the index as the first argument, as expected.