Returning 2 values from a method

This is for a mathematical method that takes in 3 doubles as input (a, b,c ).

Next those 3 values get calc开发者_StackOverflow中文版ulated in a formula (abc formula for square roots).

The formula has 2 values that have to be returned but how exactly do i do this?

I was thinking of a array / objects, but not sure how to initialise them for this purpose. Out parameters are not really of any use in this situation are they?

Regards.

In C# 4 (available in Visual Studio 2010):

Tuple<double, double> Foo(double a, double b, double c)
{
   ...
   return Tuple.Create(firstReturnValue, secondReturnValue);
}

If you're working with an earlier language version of C#, you can define your own implementation of a two-tuple (pair) as follows:

public struct Tuple<A, B>
{
   public readonly A Item1;
   public readonly B Item2;
   public Tuple(A a, B b) { Item1 = a; Item2 = b; }
}
public static class Tuple
{
   public static Tuple<A,B> Create<A,B>(A a, B b) { return new Tuple<A,B>(a,b); }
}

Of course, Tuple is a very general and unspecific data type. You could just as well implement a composite return type (similar to what I just showed, but simpler) that is more specific to your situation.

Use out:

double t;
int y = Foo(5,3,out t);

def:

public int Foo(int one,int two,out double three) {
  three = one / two;
  return one + two;
}

http://msdn.microsoft.com/en-us/library/ee332485.aspx

In C# 4 , you can do same as Tuple as described by staks above

but below c#4 , you can return a list which contains as many as values you want to return.

I'd create a class for the purpose, of even a simple struct, but a Tuple (like stakx suggested) is just as good, and even a list or a collection can do the trick.

An array is not very favorable, since you just have a bunch of items, without any further information about them, really. The caller has to somehow "know" what element no. 0, element no. 1 is.

In such a case, I would always vote for using a class - define a return type as a class, give the individual fields meaningful names. A class gives you discoverability (through field names), the ability to mix the return types (return an INT, two DECIMAL, a STRING), and it nicely encapsulates your results:

public class ResultType
{
    decimal Value { get; set; }
    decimal AnotherValue { get; set; }
    int StatusCode { get; set; } 
    string Message { get; set; }
}

and then define your method as:

ResultType YourMethodName(double a, double b, double c) 
{
     ResultType result = new ResultType();

     // do your calculations here......
     result.Value = .........
     result.AnotherValue = ............
     result.StatusCode = 0;
     result.Message = "everything OK";

     return result;
}