Question

Im using opensource library called wxFreeChart to draw some XY charts. In example there is code which uses static array as a serie :

double data1[][2] = {
                { 10, 20, },
                { 13, 16, },
                { 7, 30, },
                { 15, 34, },
                { 25, 4, },
        };

dataset->AddSerie((double *) data1, WXSIZEOF(dynamicArray));

WXSIZEOF ismacro defined like: sizeof(array)/sizeof(array[0])

In this case everything works great but in my program Im using dynamic arrays (according to users input).

I made a test and wrotecode like below:

double **dynamicArray = NULL;
        dynamicArray = new double *[5] ;

        for( int i = 0 ; i < 5 ; i++ )
            dynamicArray[i] = new double[2];

        dynamicArray [0][0] = 10;
        dynamicArray [0][1] = 20;
        dynamicArray [1][0] = 13;
        dynamicArray [1][1] = 16;
        dynamicArray [2][0] = 7;
        dynamicArray [2][1] = 30;
        dynamicArray [3][0] = 15;
        dynamicArray [3][1] = 34;
        dynamicArray [4][0] = 25;
        dynamicArray [4][1] = 4;

dataset->AddSerie((double *) *dynamicArray, WXSIZEOF(dynamicArray));

But it doesnt work correctly. I mean point arent drawn. I wonder if there is any possibility that I can “cheat” that method and give it dynamic array in way it understands it and will read data from correct place

thanks for help

Answer 1

If You define an array like

double myArr[5][2];

All cells occupy a continuous chunk of memory and I’m pretty sure dataset->AddSerie relies on that.

You can’t guarantee that if you allocate memory in chunks, using consecutive calls to new.

My proposition is to write a simple class that allocates a continuous chunk of memory for storage and uses operator() to access that memory as a two dimensional array using 2 indices. Internally You can use a vector<double> to manage the storage, and You can pass the address of the first element of that vector to dataset->AddSerie

Please check the code in this C++ FAQ example and try to understand it. The matrix example uses new[] and delete[]. You should use a vector instead, and the type double instead of Fred

Where in the example, there is a private section like this

class Matrix {
 public:
   ...
 private:
   unsigned nrows_, ncols_;
   Fred* data_;
 };

(The example shows a matrix of Freds) You should use a vector<double>

class Matrix {
 public:
   ...
 private:
   unsigned nrows_, ncols_;
   vector<double> data_;
 };

That will make the code much simpler. You don’t even need a destructor, because the vector manages the memory.