F Sharp Generics
Reversal of a list of any type[edit | edit source]
To reverse a list, it isn't important what type the list elements are, only what order they're in. This is a perfect candidate for a generic function, so the same reverseal function can be used no matter what list is passed.
let rev list = let rec loop acc = function |  -> acc | head :: tail -> loop (head :: acc) tail loop  list
The code makes no assumptions about the types of the elements. The compiler (or F# interactive) will tell you that the type signature of this function is
'T list -> 'T list. The
'T tells you that it's a generic type with no constraints. You may also see
'a instead of
'T - the letter is unimportant because it's only a generic placeholder. We can pass an
int list or a
string list, and both will work successfully, returning an
int list or a
string list respectively.
For example, in F# interactive:
> let rev list = ... val it : 'T list -> 'T list > rev [1; 2; 3; 4];; val it : int list = [4; 3; 2; 1] > rev ["one", "two", "three"];; val it : string list = ["three", "two", "one"]
Mapping a list into a different type[edit | edit source]
let map f list = let rec loop acc = function |  -> List.rev acc | head :: tail -> loop (f head :: acc) tail loop  list
The signature of this function is
('a -> 'b) -> 'a list -> 'b list, which is the most generic it can be. This does not prevent
'a from being the same type as being
'b, but it also allows them to be different. Here you can see that the
'a type that is the parameter to the function
f must match the type of the
list parameter. This function is still generic, but there are some slight constraints on the inputs - if the types don't match, there will be a compile error.
> let map f list = ... val it : ('a -> 'b) -> 'a list -> 'b list > map (fun x -> float x * 1.5) [1; 2; 3; 4];; val it : float list = [1.5; 3.0; 4.5; 6.0] > map (sprintf "abc%.1f") [1.5; 3.0; 4.5; 6.0];; val it : string list = ["abc1.5"; "abc3.0"; "abc4.5"; "abc6.0"] > map (fun x -> x + 1) [1.0; 2.0; 3.0];; error FS0001: The type 'float' does not match the type 'int'