Applying STL to Legacy Arrays

Introduction

This article shows how to apply STL algorithms to legacy arrays in a way that doesn't require much additional work. By this I mean in a way that doesn't involve copying data from the legacy arrays into an STL vector, or creating custom iterators. "Legacy" in this context refers to the built-in array or MFC's CArray.

Motivation

The motivation for this was some recent C++ maintenance programming I had to do. This code made heavy use of built-in arrays and one or two uses of the MFC CArray. When I first started using STL algorithms I wasn't aware that you could apply them to anything other than STL data structures. This is because my initial source of examples on STL was Visual C++'s online help, which always used STL data structures to illustrate the concepts. Later on I got myself Bjarne Stroustrup's The C++ Programming Language, which provided additional help and then Nicolai Josuttis' The C++ Standard Library. Both of these books explain how you can apply algorithms to built-in arrays but they don't spend much time on it and I more-or-less missed the examples first time around.

Also, I think I'd been using the basic STL data structures with their own member functions for some time before I learnt how to use the general-purpose algorithms. And, of course, since built-in arrays don't have member functions the question of manipulating them didn't arise. When I finally discovered algorithms my primary focus was in applying them to newly written code. In that context I was pretty much always going to start off with an STL data structure so, again, the issue of built-in arrays did not arise.

Built-in Arrays

Here's how to apply simple algorithms to a built-in array.

#include <iostream>
#include <algorithm>

using namespace std;

// Initialize

int a[] = { 1, 2, 1, 4, 0, 2 };
int count = sizeof(a) / sizeof(a[0]);

cout << "Initial elements\n";

for (int i = 0; i < count; i++)
{
    cout << a[i] << "\t";
}

cout << "\n";

// Find maximum element

int* begin = a;
int* end = a + count;

int* maximum = max_element(begin, end);
cout << "Maximum element = " << *maximum << "\n";

// Sort elements

sort(begin, end);

cout << "Sorted elements\n";

for (i = 0; i < count; i++)
{
    cout << a[i] << "\t";
}

cout << "\n";

This outputs:

Initial elements
1 2 1 4 0 2 
Maximum element = 4
Sorted elements
0 1 1 2 2 4 

MFC CArray

You can apply algorithms quite easily to MFC's CArray class too. There is a function, GetData(), that returns a pointer to the first array element in the CArray. Try adding the following code to, say, the OnInitDialog() member in the dialog class of an MFC dialog-based application and run it in the debugger.

#include <afxtempl.h>
#include <algorithm>

using namespace std;
    
// Initialize
    
CArray<int, int> a;
a.Add(1);
a.Add(2);
a.Add(1);
a.Add(4);
a.Add(0);
a.Add(2);

int count = a.GetSize();

afxDump << "Initial elements\n";

for (int i = 0; i < count; i++)
{
    afxDump << a[i] << "\t";
}

afxDump << "\n";

// Find maximum element

int* begin = a.GetData();
int* end = begin + count;
int* maximum = max_element(begin, end);

afxDump << "Maximum element = " << *maximum << "\n";

// Sort elements

sort(begin, end);

afxDump << "Sorted elements\n";

for (i = 0; i < count; i++)
{
    afxDump << a[i] << "\t";
}

afxDump << "\n";

This outputs:

Initial elements
1 2 1 4 0 2 
Maximum element = 4
Sorted elements
0 1 1 2 2 4 

We cannot apply this technique to a CList. For this we need to create a custom iterator. And in fact we ought to do the same for a CArray. I believe there is an example of how to do this somewhere on the CodeGuru web site: http://www.codeguru.com/. But the above technique provides a quick and dirty way of applying the power of STL in a minimally invasive way.