Today’s challenge for me was about using reduce:

defmodule Prob5 do
  def move([[h1 | rest] = _list1, list2]) do
    [rest, [h1 | list2]]
  end

  def move_n([l1, l2], n) do
    Enum.reduce(1..n, [l1, l2], fn _, acc -> move(acc) end)
  end

  def single_instruction(input, [n, from_ind, to_ind], move_function) do
    from_ind = from_ind - 1
    to_ind = to_ind - 1
    [from, to] = move_function.([Enum.at(input, from_ind), Enum.at(input, to_ind)], n)
    input
    |> List.replace_at(from_ind, from)
    |> List.replace_at(to_ind, to)
  end

  def parse_line(line) do
    [[n], [from_ind], [to_ind]] = Regex.scan(~r/\d+/, line)
    Enum.map([n, from_ind, to_ind], &String.to_integer/1)
  end

  def move_n_part2([l1, l2], n) do
    to_move = Enum.take(l1, n)
    [Enum.drop(l1, n), to_move ++ l2]
  end
end

I manually parsed the first part of the input into lists, like thus:
test_input = [[:n, :z], [:d, :c, :m], [:p]]

And then parsed the 2nd part of the input into simple lists, and finally ran it through reduce:

Enum.reduce(
  instructions, 
  input, 
  fn instruction, input -> Prob5.single_instruction(input, instruction, &Prob5.move_n/2) end)

… using the 2 move functions - one for each part.

Let me know how this can be improved!

Thanks