Matrix with variables in C

Question

I'm going to write in pseudo code to make my question more clear. Please keep in mind this code will be done in C.

Imagine I have an array of any amount of numbers. The first number tells me how big of an array we're dealing with. For example, if my first number is 3, it means I have two 3x3 matrices. So I create two multidimensional arrays with:

matrix1[3][3]
matrix2[3][3]

What I'm having a hard time with is the arithmetic/coding to assign all the numbers to the matrices, I'm having a very hard time visualizing how it would be done.

Imagine a test array contains [2,1,2,3,4,5,6,7,8]

My program should now have two matrixes with:

1 2    5 6
3 4    7 8

Do I need several nested loops? Any help would be appreciated.

Answer 1

Your data is presented in row-major order. After reading your integer array and validating the content (i.e. the dim=4 means 32 values follow, dim=2 means 8 values follow, etc.) I'm not sure why you want to allocate or loop anything.

I.e. You can use your physical test[] data as the matrices:

int dim = test[0];
int (*mat1)[dim] = (int (*)[dim])(test+1);
int (*mat2)[dim] = (int (*)[dim])(test+1 + dim*dim);

C99 supports variable array declarations at the implementation level (i.e. the compiler can support the feature as it is defined by the standard, but does not have to; see 6.7.6.2 of the C99 standard for more info). If your toolchain does NOT support it, then a predefined macro, __STDC_NO_VLA__ must be defined and can be tested at compile time (see section 6.10.8.3-1 of the C99 standard). That being said, every C99-compliant compiler I've ever used in the last decade-plus does support it, so if your's does not, tell us below in a comment.

If it does, then pay note to the use of 'dim' in the declarations of mat1 and mat2 above). It is one of the few features of C I like that C++ does not have. So dance with the one you brought.

Finally, assuming your compiler is C99 compliant and supports VLAs (__STDC_NO_VLA__ is NOT defined), as an extra super-special bonus it is all-but-guaranteed to be the fastest algorithm to get your two matrices, because there is no algorithm. You read one array element, then assign two pointers. O(3) is hard to beat.

Example

#include <stdlib.h>
#include <stdio.h>

// main loader.
int main(int argc, char *argv[])
{
    int test[] = {2,1,2,3,4,5,6,7,8};
    int dim = test[0];
    int (*mat1)[dim] = (int (*)[dim])(test+1);
    int (*mat2)[dim] = (int (*)[dim])(test+1 + dim*dim);

    // proof stuff is where it should be.
    int i=0,j=0;
    for (i=0;i<dim;i++)
    {
        for (j=0;j<dim;printf("%d ", mat1[i][j++]));
        printf ("   ");
        for (j=0;j<dim;printf("%d ", mat2[i][j++]));
        printf("\n");
    }

    return EXIT_SUCCESS;
}

Output

1 2    5 6 
3 4    7 8 

A similar test with a 3x3 data set:

int test[] = {3,1,2,3,4,5,6,7,8,9,9,8,7,6,5,4,3,2,1};

Output

1 2 3    9 8 7 
4 5 6    6 5 4 
7 8 9    3 2 1 

And finally, a 4x4 data set:

int test[] = {4,1,2,3,4,5,6,7,8,1,2,3,4,5,6,7,8,8,7,6,5,4,3,2,1,8,7,6,5,4,3,2,1};

Output

1 2 3 4    8 7 6 5 
5 6 7 8    4 3 2 1 
1 2 3 4    8 7 6 5 
5 6 7 8    4 3 2 1 

Answer 2

Here is something that works in a single loop; no nests and no repeats. I don't know if it'll outperform other answers, but I just felt like giving you a different answer ^_^

I have not tested this code, but it looks like the logic of the algorithm works - that's the point, right? Let me know if it has any errors....

int c=0, x=0, y=0, size=test[0], length=sizeof(test);

for(i=1; i<length; i++) {
    if((c-size)<0) {
        matrix1[x][y] = test[i];
    } else {
        matrix2[x][y] = test[i];
    }
    ++y;
    if(y%size == 0) {
        ++c;
        y = 0;
        x = (c-size)<0 ? ++x : 0;
    }
}

Answer 3

2D arrays are stored in continuous series of lines from the matrix. So you doesn't even need to allocate new memory you can use your original array. Anyway you can create 2 new standalone array too. You can crate a function like this, to get the correct number of the matrix.

int getNumber(int array[], int arraynumber, int index_x, int index_y)
{
   return array[(((array[0]*index_x)+index_y)+1)+((array[0]*array[0])*arraynumber)];
}

The arraynumber variable is 0 for the first and 1 for the second matrix. This funciton works only if all parameters are correct, so ther is no error detection.
With this function you can easily loop through and create 2 new arrays:

int i,k;
for (i=0; i<array[0]; i++)
{
   for (k=0; k<array[0]; k++)
   {
      newarray1[i][k] = getNumber(array, 0, i,k);
      newarray2[i][k] = getNumber(array, 1, i,k);
   }
}

Answer 4

The problem with multidimensional arrays in C is that you need to know in advance (at compile time) n-1 of the dimension sizes, they are also a drag when used as function parameters.

There are a couple of alternate approaches:

1. Creating an array of arrays. i.e. allocating an array of array pointers and then allocating arrays to those pointers.

   type **array = malloc(sizeof(type * ) * < firstnumread > );
   array[0] = malloc(sizeof(type) * < firstnumread > );
   ...

2. Allocating a single dimension array with the size of all the multiplied dimensions. i.e.

   type *array = malloc(sizeof(type) * < firstnumread > * < firstnumread >);

In your case, the second is probably more appropiate. Something like:

matrix1 = malloc(sizeof(type)*<firstnumread>*<firstnumread>);  
matrix2 = malloc(sizeof(type)*<firstnumread>*<firstnumread>);

Then you can assign values like this:

matrix1[column*<firstnumread> + row] = <value>;

Yes, with 2 for loops.

Answer 5

At the moment the only idea i get is using two for loops. Or you can make a function and call it every time you need (but don't forget to use k as second argument).

int i, j, k;

/* We start in the 2nd element of the array that's why k = 1. */
k = 1;

/* Now we fill the array1 copying 1 by 1 the elements of the "test array" until
   we fill it. Then we do the same with the array2. */
for( i = 0; i < test[ 0 ]; i++ ){
   for( j = 0; j < test[ 0 ]; j++ ){
       array1[ i ][ j ] = test[ k ]
       k++;
   }
}

for( i = 0; i < test[ 0 ]; i++ ){
   for( j = 0; j < test[ 0 ]; j++ ){
       array2[ i ][ j ] = test[ k ]
       k++;
   }
}