I would like to make a function that accepts a list and returns two lists: the first contains every odd item, and the second contains every even item.
For example, given [1;2;4;6;7;9], I would like to return [ [1;4;7] ; [2;6;9] ].
I have written this so far and I do not know how to progress.
let splitList list =
let rec splitOdd oList list1 list2 =
match oList with
| [] -> []
| head :: tail -> splitEven tail (list1::head) list2
and splitEven oList list1 list2 =
match oList with
| [] -> []
| head :: tail -> splitOdd tail list1 (list2::head)
splitOdd list [] []
Implementation which does not stack-overflows:
let splitList list = List.foldBack (fun x (l,r) -> x::r, l) list ([],[])
If you mean odd and even values for positions of items, here is a (non-tail-recursive) solution:
let rec splitList = function
| [] -> [], []
| [x]-> [x], []
| x1::x2::xs -> let xs1, xs2 = splitList xs
x1::xs1, x2::xs2
Here is a straightforward non-recursive solution:
let splitList ll =
ll
|> List.mapi (fun i x -> (i % 2 = 0, x))
|> List.partition fst
|> fun (odd,even) -> [List.map snd odd, List.map snd even];;
val splitList : 'a list -> 'a list list
Being applied to your sample it yields exactly what you wanted:
splitList [1;2;4;6;7;9];;
val it : int list list = [[1; 4; 7]; [2; 6; 9]]
It looks like this is what you were going for, which is indeed a nice way to do it as it's tail-recursive.
let splitList items =
let rec splitOdd odds evens = function
| [] -> odds, evens
| h::t -> splitEven (h::odds) evens t
and splitEven odds evens = function
| [] -> odds, evens
| h::t -> splitOdd odds (h::evens) t
let odds, evens = splitOdd [] [] items
List.rev odds, List.rev evens
Another (less efficient) option
let splitList xs =
let odd, even =
xs
|> List.zip [ 1 .. (List.length xs) ]
|> List.partition (fun (i, _) -> i % 2 <> 0)
[ odd |> List.map snd; even |> List.map snd ]
If you want to avoid creating temporary lists, consider using sequences:
let splitListSeq xs =
xs
|> Seq.mapi (fun i x -> (i % 2 = 0, x))
|> Seq.groupBy (fun (b, _) -> b)
|> Seq.map snd
|> Seq.map ((Seq.map snd) >> Seq.toList)
|> Seq.toList
Yet, another one, similar to Daniel's version:
let splitListRec xs =
let rec loop l r = function
| [] -> [l; r]
| x::[] -> [x::l; r]
| x::y::t -> loop (x::l) (y::r) t
loop [] [] xs |> List.map List.rev
Sounds like you want List.partition<'T> ( http://msdn.microsoft.com/en-us/library/ee353782.aspx ). This function takes a predicate and a list and will return a pair (2-tuple) where the first element is all the elements that passed your test and the second is all the elements that did not pass the test. So you could classify odds and evens with:
List.partition odd [1;2;4;6;7;9]
If your really want a list, you can use fst and snd to extract the elements from the tuple and put them in a list.
Good luck!
My 2¢, in OCaml, since there still is a bounty open.
Maybe you could give us a hint what you want. Elegance? FP? Tail recursion? Performance?
Edit:
I removed the longer solution. For List.partition to work, the predicate was missing. Here it is:
let so_split lst =
let flag = ref false in
List.partition (fun e -> flag := not !flag; !flag) lst
Improvements any? Testing the solution:
# so_split [1;2;4;6;7;9];;
- : int list * int list = ([1; 4; 7], [2; 6; 9])
Just for completeness here is a boring, more imperative solution:
let splitList (list:int list) =
let odds = [for i in 0..list.Length-1 do
if i%2=1 then
yield list.[i]]
let evens = [for i in 0..list.Length-1 do
if i%2=0 then
yield list.[i]]
odds,evens
来源:https://stackoverflow.com/questions/7942630/splitting-a-list-of-items-into-two-lists-of-odd-and-even-indexed-items