Real-world examples where C# 'out' parameters are useful?

Question

I\'m reading up on core C# programming constructs and having a hard time wrapping my head around the out parameter modifier. I know what it does by reading but

Answer 1

Output parameters are found all over the .NET framework. Some of the uses I see most often are the TryParse methods, which return a boolean (indicating whether or not the parse was valid) and the actual result is returned via the output parameter. While it's also very common place to use a class when you need to return multiple values, in such an example as this it's a little heavy handed. For more on output parameters, see Jon Skeet's article on Parameter passing in C#.

Answer 2

bool Int32.TryParse(String, out Int);

or something similar like Dictionary.TryGetValue.

But I would consider this one to be a not too good practice to employ it, of course, using those provided by API like the Int32 one to avoid Try-Catch is exceptions.

Answer 3

//out key word is used in function instead of return. we can use multiple parameters by using out key word
public void outKeyword(out string Firstname, out string SecondName)
{
    Firstname = "Muhammad";
    SecondName = "Ismail";

}
//on button click Event
protected void btnOutKeyword_Click(object sender, EventArgs e)
{
    string first, second;
    outKeyword(out first, out second);
    lblOutKeyword.Text = first + "  " + second;
}

Answer 4

Sure, take a look at any of the TryParse methods, such as int.TryParse:

The idea is you actually want two pieces of information: whether a parse operation was successful (the return value), and, if so, what the result of it actually was (the out parameter).

Usage:

string input = Console.ReadLine();
int value;

// First we check the return value, which is a bool
// indicating success or failure.
if (int.TryParse(input, out value))
{
    // On success, we also use the value that was parsed.
    Console.WriteLine(
        "You entered the number {0}, which is {1}.",
        value,
        value % 2 == 0 ? "even" : "odd"
    );
}
else
{
    // Generally, on failure, the value of an out parameter
    // will simply be the default value for the parameter's
    // type (e.g., default(int) == 0). In this scenario you
    // aren't expected to use it.
    Console.WriteLine(
        "You entered '{0}', which is not a valid integer.",
        input
    );
}

Many developers complain of out parameters as a "code smell"; but they can be by far the most appropriate choice in many scenarios. One very important modern example would be multithreaded code; often an out parameter is necessary to permit "atomic" operations where a return value would not suffice.

Consider for example Monitor.TryEnter(object, ref bool), which acquires a lock and sets a bool atomically, something that wouldn't be possible via a return value alone since the lock acquisition would necessarily happen before the return value were assigned to a bool variable. (Yes, technically ref and out are not the same; but they're very close).

Another good example would be some of the methods available to the collection classes in the System.Collections.Concurrent namespace new to .NET 4.0; these provide similarly thread-safe operations such as ConcurrentQueue<T>.TryDequeue(out T) and ConcurrentDictionary<TKey, TValue>.TryRemove(TKey, out TValue).

Answer 5

The main motivation to using an out parameter is to allow a function to return multiple values to the caller and everyone else provided examples in the framework. I'll take a different approach to answering your question by exploring the reasoning behind having out parameters in the first place. I won't write out actual examples but describe them.

Normally you have only one mechanism to return values, the function's return value. Sure you could use a global (static) or instance variables too but that's not very practical nor safe to do in general (for reasons I won't explain here). Prior to .NET 3.5, there wasn't a really practical way to return multiple values from a function. If out or ref modifiers were not available, you would have a few options:

• If all your values had the same type, you could return some collection of the values. This is perfectly fine in most cases, you could return an array of number, list of strings, whatever. This is perfect if all the values were related in exactly the same way. i.e., All numbers were the number of items in a container, or the list was of names of guests at a party. But what if the values you returned represented different quantities? What if they had different types? A list of objects could hold them all but it is not a very intuitive way to manipulate that sort of data.

• For the case when you need to return multiple values of different types, the only practical option you had was to create a new class/struct type to encapsulate all these values and return an instance of that type. Doing so you could return strongly typed values with intuitive names and you could return multiple values this way. The problem is that in order to get that, you had to define the type with a specific name and everything just to be able to return multiple values. What if you wanted to return only two values which were simple enough making it impractical to create a type for it? You have a couple more options again:

  • You could create a set of generic types to contain a fixed amount of values of varying types (like a tuple in functional languages). But it is not as appealing to do so in a reusable manner since it wasn't part of the framework at the time. It could be put in a library but now you add a dependency on that library just for the sake of these simple types. (just be glad that .NET 4.0 now includes the Tuple type) But this still doesn't solve the fact that these are simple values which means added complexity for a simple task.

  • The option that was used was to include an out modifier which allows the caller to pass a "reference" to a variable so that the function may set the referenced variable as another way to return a value. This way of returning values was also available in C and C++ in many ways for the same reasons and played a role in influencing this decision. However the difference in C# is that for an out parameter, the function must set the value to something. If it doesn't, it results in a compiler error. This makes this less error prone since by having an out parameter, you're promising the caller that you will set the value to something and they can use it, the compiler makes sure you stick to that promise.

A note on the typical usage of the out (or ref) modifier, it will be rare to see more than one or two out parameters. In those cases, it will almost always be a much better idea to create the encapsulating type. You would typically use it if you needed just one more value to return.

However since C#-3.0/.NET-3.5 with the introduction of anonymous types and tuples introduced in .NET 4.0, these options provided alternative methods to return multiple values of varying types easier (and more intuitive) to do.

Answer 6

there are many scenarios where you would use it, but the main one would be where your method needs to return more then one parameter. Take, for example, the TryParse methods on int type. In this case, instead of throwing an exception a bool is returned as a success/failure flag and the parsed int is return as the out param. if you were to call int.Parse(...) you could potentially throw an exception.

string str = "123456";
int val;
if ( !int.TryParse(str,out val) )
{
// do some error handling, notify user, etc.
}