Non Square Matrix Multiplication Help C++ - c++

Question

first off sorry for the slightly messy code as I was fiddling with different things to try to get it to work. As of now, my code can multiply Square Matrices just fine; however, it has difficulties computing Non Square Matrices. My best guess after debugging is how i re-size my vectors, and that there is an out of bounds error which causes the program to crash. Any help would be appreciated, my code should be able to Multiply any Vectors sizes withing Matrix Multiplication Rules.
I also would like to note this is a HW assignment so I am limited to how I can build my code, basically ONLY using vectors, cannot write your own class etc....
#include <iostream>
#include <vector>
using namespace std;
void multiply_matrices(vector <vector<int> > matrix1,vector <vector<int> > matrix2, int cols, int rows2);
void setMatrix(vector <vector<int> > &matrix, int rows, int cols);
int main()
{
int rows, cols, rows2, cols2;
vector< vector<int> > matrix, matrix2;
cout<<"Please enter the number of Rows and Columns for your first Matrix."<<endl;
cout<<"Rows: ";
cin>>rows;
cout<<"Columns: ";
cin>>cols;
matrix.resize(cols, vector<int>(rows,0)); //Saw this online so not sure how it works but it works, if i take out one i cant do row<column and vice versa
matrix.resize(rows, vector<int>(cols,0));
cout<<"Size has been declared, please enter data for your matrix"<<endl;
setMatrix(matrix,rows,cols);
cout<<"Second Matrix Automatically Set by Matrix Multiplication Rule"<<endl; //Just automatically sets second matrix as per Matrix Multiplication Rule
rows2=cols;
cols2=rows;
cout<<"Second Matrix Size is: " << rows2 << " by " << cols2 << endl;
matrix2.resize(cols2, vector<int>(rows2,0));
matrix2.resize(rows2, vector<int>(cols2,0));
setMatrix(matrix2,rows2,cols2);
cout<<"Multiplied Matrix is:"<<endl;
multiply_matrices(matrix,matrix2,cols,rows2);
system("PAUSE");
return 0;
}
void setMatrix(vector <vector<int> > &matrix, int rows,int cols){
int num;
for(int i = 0; i < rows; i ++)
{
for (int j = 0; j < cols; j++)
{
cout << "Enter Value for Row " << (i+1) << " Column " << (j+1) << ": ";
cin>>num;
matrix[i][j]=num;
}
cout << endl;
}
/*for(int i = 0; i < rows; i ++)
{
for (int j = 0; j < cols; j++)
{
cout << matrix[i][j] << " ";
}
cout << endl;
}
*/
}
void multiply_matrices(vector <vector<int> > matrix1,vector <vector<int> > matrix2, int cols, int rows2){
vector< vector<int> > tempMatrix;
int newrows=rows2;
int newcols=cols;
int sum;
tempMatrix.resize(newcols, vector<int>(newrows,0)); //Resizing new matrix to proper size, so if it was (2x3)(3x2), new matrix is (3x3)
for (int i = 0; i < newrows; i++) //This Works Fine for Square Matrixes but not for others, i have no clue how to fix it?
{
for (int j = 0; j < newcols; j++){
//sum=0;
for (int u = 0; u < newcols; u++)
{
//sum+=matrix1[i][u] * matrix2[u][j];
//tempMatrix[i][j]=sum;
tempMatrix[i][j] += matrix1[i][u] * matrix2[u][j];
}
}
}
for(int i = 0; i < newrows; i ++)
{
for (int j = 0; j < newcols; j++)
{
cout << tempMatrix[i][j] << " ";
}
cout << endl;
}
}

Answer 1

Initialize your first matrix like this:
matrix.resize(rows, vector<int>(cols,0));
and your second like this:
matrix2.resize(rows2, vector<int>(cols2,0));
where rows2 = cols. Note that there is no "multiplication rule" that implies cols2 == rows.
The problem is in your multiply_matrices function where the loops should be
for (int i = 0; i < rows; i++) // or matrix1.size()
for (int j = 0; j < cols2; j++) // or tempMatrix[i].size()
for (int u = 0; u < cols; u++) // or rows2 or matrix1[i].size()
but as already stated in the comments, it would be better to use vector::size() instead of passing the sizes as additional parameters.
Also, if you multiply (2x3)(3x2), the result is (2x2):
tempMatrix.resize(rows, vector<int>(cols2,0));

Answer 2

There is nothing wrong with the resize() function. What may be wrong is that you ignore the maximum size and rely solely on variables that are passed to your functions.
For example, your setMatrix function is passed rows and cols, but this is not necessary.
The function should be rewritten using only the matrix to provide the sizes to the loops:
void setMatrix(vector<vector<int> > &matrix)
{
int num;
for(int i = 0; i < matrix.size(); ++i)
{
for (int j = 0; j < matrix[i].size(); ++j)
{
cout << "Enter Value for Row " << (i+1) << " Column " << (j+1) << ": ";
cin>>num;
matrix[i][j] = num;
}
cout << endl;
}
}
You have the same issue with multiply_matrix. What you should be doing is ensure that you're loops use the return value of vector::size(), which you do not do. The problem is here:
for (int i = 0; i < newrows; i++)
{
for (int j = 0; j < newcols; j++)
{
for (int u = 0; u < newcols; u++)
{
tempMatrix[i][j] += matrix1[i][u] * matrix2[u][j];
You sized the tempMatrix to newrows rows and newcols columns. But how do you know that matrix1 and matrix2 have at least newrows rows and newcols columns? You don't know, but you just assume they do.
So you either need to ensure that the size of matrix1 and matrix2 can accommodate the number of rows/columns, or you throttle those loops to use the minimum rows/columns.
Overall, the issue is that nowhere do you use vector::size() in your code that I see. So start to use size() to your advantage -- don't create superfluous (and possibly erroneously set) variables that supposedly denote the sizes of the rows and columns.