OCaml Parsing String into String*Int list

Question

type unit_t = (string * int) list

How would I go about creating an expression to parse a string into this form? Do I need to implement code in a parser.mly format in order to handle this?

Answer 1

There's no real way to give a detailed answer, since you don't say how the original string is supposed to correspond to your list of pairs.

Based just on the simplicity of your output type, I'd say you could do this without invoking the full machinery of ocamlyacc. I.e., I doubt you need a parser.mly file. You could use ocamllex, or you could just apply some regular expressions from the Str module. It depends on what your parsing actually needs to do.

Here's one possibility that uses regular expressions from Str:

let p s =
    let fs = Str.full_split (Str.regexp "[0-9]+") s in
    let rec go = function
    | [] -> []
    | Str.Text s :: Str.Delim i :: rest -> (s, int_of_string i) :: go rest
    | Str.Text s :: rest -> (s, 0) :: go rest
    | Str.Delim i :: rest -> ("", int_of_string i) :: go rest
    in
    go fs

You can run it like this:

# #load "str.cma";;
# #use "p.ml";;
val p : string -> (string * int) list = <fun>
# p "123abc456def";;
- : (string * int) list = [("", 123); ("abc", 456); ("def", 0)]
# p "ghi789jkl100";;
- : (string * int) list = [("ghi", 789); ("jkl", 100)]