C++: Functions

Looking for in-depth C tutorials? Try SourceCrunch.

We've talked a little bit about functions in the past - they're pieces of code that can be executed on command. In fact we've been using functions right since the very start of this process, we just haven't really talked about them in depth - this is what this tutorial is all about.

As outlined earlier - a function is simply a piece of code (which may or may not need a value to work properly) that you can execute by "calling" it. There are a bunch of functions built into the core language, but functions can also be created manually (we'll be learning about this a bit later on). The classic function notation looks like this: functionName();. In the previous snippet we would be calling a function called functionName which doesn't take any values to work properly (if it took any values - we'd put them in the brackets). A classic example of a function is the sqrt function which is defined inside math.h. We used this back in the basic mathematics tutorial, it simply square roots any value you pass to it. If we ever need to pass multiple values to a function, we separate them with a comma - so something like this: functionName(valueone, valuetwo, valuethree);. With this under our belts - let's start learning how to create our very own functions that can execute any preset code for us. For this tutorial we are going to work on creating a function which can add two numbers together.

Believe it or not, you've already got one function in your basic program structure. The main function! Let's dissect it a little so we can use something similar to create custom functions. The first thing it consists of is the function's type - it's an int in this case because the main function needs to be an int to be the int main that the compiler recognises as the program's entry point. The next thing (seperated from the type by a space) is the function's name - in this case it's main, and as I just alluded to, this is so the compiler recognises it as a point of entry. Next we have some empty brackets - these are here to illustrate that the function doesn't take any values, if we wanted a function to take values in (in the case of the function we want to create - we'll need to take two values so we can add them together) then we would put them inside the brackets. From here we simply have some curly brackets, inside which is the actual code we want the function to execute when called.

So firstly - why do functions need a type? Well, it's all to do with return. If a function is of type int, it must return an integer (e.g. return 0 which in the case of the main function illustrates that the program exited without any errors). Values that functions return essentially go in the place where the function was originally called - so the sqrt function returns the result of the square root operation (hence when we put it in a cout or something, the square root value get's inserted at the point we called it). Function can have the type of any data-type (or struct - we'll learn about these way later in the course) and can also have the type of void (which means they don't need to return anything).

So before we actually start creating the "proper" version of our function, let's just create a function which can output some basic text - let's stick with the classics and make it output "Hello World". So firstly let's decide on a type and name - void is probably a good choice here as we can just use cout in the function rather than making it return anything. For the name, I'm going to go with "say_hello". So to make the function, we simply write a definition (the structure of which we've already talked about) before the main function in the code. It's important that it's before, otherwise the compiler won't know what you're talking about when you reference the function in main. The following would work fine for our simply say_hello function (I've just used the basic structure we've already discussed and put a cout inside):

1
2
3
4
void say_hello()
{ //I like dropping the curly brackets to separate lines, it's personal formatting preference
	cout << "Hello World";
}

Alternatively, you could put the function definition below the main function and simply leave what is called a prototype before the main function (to tell the compiler not to worry when trying to call the function in main as we define it later). A prototype is essentially just the first line of our function definition (no curly brackets) with a semi-colon afterwards. So if we wanted to use the prototype approach instead of putting the function before (some people prefer doing things this way - a lot of it is down to personal preference), we could do something like this:

1
2
3
4
5
6
7
8
9
10
11
void say_hello(); //Our function prototype to tell the compiler not to worry

int main()
{
	//Classic stuff in here (including the function call if we actually want to use it)
}

void say_hello()
{ //I like dropping the curly brackets to separate lines, it's personal formatting preference
	cout << "Hello World";
}

Now you've got the function setup (either before the main function, or after with a function prototype before) - you can go ahead and call it from the main function using the process we talked about earlier (while we're talking about this - it's worth noting that functions can call each other if you want this nested-like functionality with multiple functions).

So our code at the moment should look something like this (I'm not going to use prototypes in this example just to avoid confusing anyone who doesn't quite understand them):

1
2
3
4
5
6
7
8
9
10
11
12
13
14
#include <iostream>

using namespace std;

void say_hello()
{
	cout << "Hello World";
}

int main()
{
	say_hello();
	return 0; //You should understand why this is necessary now :)
}

On running the above program, you should see that our basic function works! Brilliant, now let's modify it a bit so that it can add two numbers. Firstly let's change its name from "say_hello" to something more appropriate like "add_two_integers". For now we can leave it as void and then make it simply do the addition in the cout - but we'll change this later and make it instead use return. Now let's make it take in two integer values, as it requires these to do its job properly (how else would it add two numbers?) - we write these just like variable definitions apart from they are inside the function definition's brackets (and inside the prototype brackets too if you're using a function prototype). If you want the function to take more than one value (like we do), then you can separate the values with commas. I'm going to call the parameters that my function takes (the proper terminology for the values it takes), "value_one", and "value_two". Our function definition at the moment should look something like this:

1
2
3
4
void add_two_integers(int value_one, int value_two)
{
	cout << "Hello World";
}

From here, we can reference "value_one" and "value_two" (or whatever you called the "variables" in the function definition) in our function and it will refer to whatever values the function was called with. It's important to note that if you try to call the function with two values that do not match the type of those that you put in the function definition (or you specify a different number of values) - the compiler should spit out an error telling you that it can't find the function you're talking about. So with the knowledge that we can reference variables that the function was passed, let's make our function output the addition of the two variables it was passed:

1
2
3
4
void add_two_integers(int value_one, int value_two)
{
	cout << value_one+value_two;
}

It's that simple, as long as we update our function call so that it uses the correct function name and passes the correct values, everything should work!

1
2
3
4
5
6
7
8
9
10
11
12
13
14
#include <iostream>

using namespace std;

void add_two_integers(int value_one, int value_two)
{
	cout << value_one+value_two;
}

int main()
{
	add_two_integers(10, 5);
	return 0;
}

Excellent! The only problem is that our function for adding two integers isn't usable if we want to put the number in a more complex cout, do more mathematics to it, or pass it to another function. Our function's name is also a little bit misleading as the function doesn't actually add two integers, it outputs the addition of two integers. To fix this, let's just change the function type and use return. Since the addition of two integers should always result in an integer, we can use the int function type. Now we've done this we must make the function return a value or the compiler will (should) complain. Let's go ahead and remove the cout line currently in there as it's getting in the way, and replace it instead with a line that returns the addition.

1
2
3
4
int add_two_integers(int value_one, int value_two)
{
	return value_one+value_two;
}

All we need to do now is update the function call in main so we actually use the value it returns somehow - I'm just going to multiply the addition of the two values by 2. The full source code should look something like the following:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
#include <iostream>

using namespace std;

int add_two_integers(int value_one, int value_two)
{
	return value_one+value_two;
}

int main()
{
	cout << add_two_integers(10, 5)*2;
	return 0;
}

And tada, we created our very own fully functional function! This is a very important technique when creating applications/programs using C++ and should be a process you get very used to in the creation process. If you fancy the challenge, try creating a basic function that takes three parameters - it adds the first two, and then multiplies the result by the third parameter if the third parameter is less than 100.

There are a few more things we should probably talk about before this tutorial is over, namely what we can do when some parameters are not specified. The first thing we can do is something called overloading functions. This is a key concept in C++ which allows you to write two or more functions with the same name that take different parameters!

At current, our "add_two_integers" function will only be called when we pass it two integer values, however we could, if we wanted, write a function with the same name that takes different parameters or parameters of different types, and the compiler would call the correct function for the job depending on the parameters we passed in. Let's say, for example, that we also wanted a version of our function that added floats -- if we renamed the function to "add_two_numbers" (so that the name makes more sense), we could write both versions of our function (one for integers and one for floats) like so:

1
2
3
4
5
6
7
8
9
int add_two_numbers(int value_one, int value_two)
{
	return value_one+value_two;
}

float add_two_numbers(float value_one, float value_two)
{
	return value_one+value_two;
}

In the example above, we could then call "add_two_numbers" with either two integers or two floats without a problem, and the appropriate function would be called. Similarly, we could, if we wanted, make a function with the same name that just takes completely different parameters, and the compiler would choose the right one, spitting out an error if it can't find one which takes the parameters we're giving it. This concept is very important when creating bigger applications and is used all over the place in the world of C-programming.

The other way we could address the "issue" of the wrong number of parameters being passed (although not the wrong type), is by using default parameters. This is essentially where we can just give a value to use for a parameter if none are specified, so in our "add_two_numbers" function we might want to default the values to 0 and 0 if no parameters are passed to it -- this can be accomplished via an equals sign in the function parameter declaration brackets:

1
2
3
4
int add_two_numbers(int value_one = 0, int value_two = 0)
{
	return value_one+value_two;
}

In the above we could just call add_two_numbers (); with no problem. It's important to remember that after the first default parameter in a function declaration (e.g. int value_one = 0), all preceding parameters must also have default values assigned to them (and if you think about it, this makes sense).

It heavily depends on usage situations, but function overloads are often preferable to default parameters in situations in which overloads don't cause a whole bunch of code repeating -- this is because mixing default parameters and function overloads can mess things up a little bit. Consider the following in which the compiler doesn't know which function to use:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
int add_two_numbers(int value_one = 5, int value_two = 5)
{
	return value_one+value_two;
}
int add_two_numbers(int value_one)
{
	return value_one;
}

int main()
{
	add_two_numbers(5);
	return 0;
}

Your compiler should spit out an error in cases like this - something along the lines of "ambiguous call to overloaded function", but be careful as this can really mess up your code and cause a lot of confusion!

To tie off this tutorial, let's just touch on a concept called recursion. Recursion, in the context of functions, is where a function calls itself. This concept can seem a little bit confusing at first, however it shouldn't be totally alien to you. Usually it's best when people figure out recursion for themselves, so all I'm going to do to help you along is show you the following example of recursion:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
void countdown(int n)
{
	cout << n << endl;
	n--;
	if(n > 0)
	{
		countdown(n); //The function calls itself! Recursion!
	}
}

int main()
{
	countdown(10); //Count down from 10

	return 0;
}

Recursion can be a little mind-bending, and in the above example isn't entirely necessary, but it's a very important topic when programming and is the best solution in a number of situations. If you don't understand the above code snippet, try running through the program line-by-line, starting from the 'countdown' call in 'main', and just do exactly what your program would do upon execution in your mind. We've covered a lot in this tutorial so you should be doing a lot of experimentation to test out the different things demonstrated -- none of the concepts covered should be hard as nails, but you need to make sure you fully understand them before moving on to the next tutorial (as is the case with most of the tutorials in this course - each tutorial builds on things learnt from the last).