Question

I am learning Jason Hickey’s Introduction to Objective Caml.

---

It says

Matching against floating-point values is supported, but it is rarely used because of numerical issues

---

Ok, we can’t match floating-point values.

Then what if we need? How to do it then?

Answer 1

To expand on previous answers, you can use pattern matching with floats, like in this example:

# let float_match_example = function
    | 0. -> "exact zero"
    | x when abs_float x < 1e-12 -> "epsilon"
    | _ -> "other";;

val float_match_example : float -> string = <fun>

# List.map float_match_example [0.; 42e-15; 3.];;

- : string list = ["exact zero"; "epsilon"; "other"]

Just be careful that pattern matching implicitly uses equality test, which is rarely what one wants when working with floats: when rounding errors accumulate, two floats are rarely exactly equal. So, you can use guards with the keyword when, as illustrated in the above code.