Build weighted matrix in MATLAB with vectors

Question

I have a problem using MATLAB. I want to create a huge adjacency matrix and to achieve this result I wrote some code that generates 3 vectors: NodeX , NodeY , and Weight .

NodeX represents the row indexes of the matrix.

NodeY represents the column indexes of the matrix.

Weight is a vector that contains the weights.

All these vectors are put into a .csv file that MATLAB reads very easily. But when I create the matrix I have a problem. This is the code that I use to associate indexes to weights.

    for i=1:1:3873
    WeightedMatrix(NodeY(i), NodeX(i)) = Weight(i);
    end

The results I see are a little bit strange to me. In fact if I do:

WeightedMatrix(NodeY(1), NodeX(1))

MATLAB prints to me:

ans = 0.2483

Which is a value that is present into the Weight vector but it's not the value I want.

And if I do:

Weight(1)

It returns ans = 1.2550 (which is the value I want).

These three vectors are for example:

This is the final part that creates the .csv file:

tmp = set(Final)
tmp = tuple(tmp)
for z in tmp:
print str(z[0]) + "," + str(z[1]) + "," + str(z[2]) + "\n"

Answer 1

You can form adjacency matrix as a sparse matrix to prevent memory problem:

WeightedMatrix_sp = sparse(NodeX, NodeY, Weight);

If you have enough memory you can convert sparse matrix to dense matrix:

WeightedMatrix = full(WeightedMatrix_sp);