While reading Dave Thomas's book, Programming Elixir, I stumbled upon an exercise.
You're supposed to write a flatten(list) function that takes a list as a parameter. That list may contain any number of sublists, which themselves may contain sublists ... Your function should return all the elements of these lists in a single flattened list, all the while preserving the order of the elements.
In the beginning, it sounded a bit hard (especially since he mentioned in the book that the exercise it's hard), but after enough playing with the idea, I got to a fairly simple solution, simpler than what I thought it would be and the solutions I've found so far on the internet.
The result should look something like this:
iex> MyList.flatten([ 1, [ 2, 3,  ], 5, [[]]]) [1,2,3,4,5,6]
Now the function:
defmodule MyList do def flatten(), do:  def flatten([head | tail]) when is_list(head) do flatten(flatten(head) ++ tail) end def flatten([head | tail]), do: [head | flatten(tail)] end
At first I thought it wouldn't work, but tested it and got surprised.
The flatten function takes care of 3 cases:
- the given list is empty
- the head of the list is in itself a list, in which case it should be flattened
- the head is not a list, in which case we continue with flattening the tail