Using `dynamic` for arbitrary number types

Question

I keep having use cases where it would be nice to have a parent Number type that the numeric primitives inherit from, but to the best of my research, none exists. The next best solution I've found is using the dynamic keyword.

For example, if I want to write a function to double a number, I don't care if that number is an int or a double as the code will be the same: return value * 2;.

My best approximation of this would be using dynamic to simply resolve types at runtime and using a switch statement to filter types via pattern matching


static void Main(string[] args)
{
    double numberD = DoubleNumber(Math.PI);
    float numberF = DoubleNumber((float)Math.PI);
    int numberI = DoubleNumber((int)Math.PI);
}

static T DoubleNumber<T>(T value)
{
    return value switch
    {
        int or float or double => (dynamic)value * 2,
        _ => throw new InvalidCastException(typeof(T).ToString()),
    };
}

However, this is still jank since at compile time, literally any type could be sent as input to this function. If a string is passed into the function, that will not be caught until runtime.

Is there a way to rewrite my function signature such that it will only take int, uint, float, or double, without copy/pasting the backing code into separate functions targeting these types individually?

A coworker helped me figure this out. What I needed was the `INumber` interface. My question was not duplicate. — EagleBirdman, Feb 08 '23 at 15:11
meh, the *intent* is a duplicate; the exact words: sure, but it is the *intent* that matters, mostly — Marc Gravell, Feb 08 '23 at 15:41

Using `dynamic` for arbitrary number types

0 Answers0