Attached below is the code with the following changes:
- Matrix dimension is nxn and not hardcoded.
- Nested for-loops are removed by using vectorized operations. This increases readability as well as performance of code. You can read about vectorized operations here Vectorization - MATLAB & Simulink - MathWorks India
n = sqrt(size(A, 2) - 1); % Assume A has 1 column for ID and n^2 columns for matrix values
for iter = 1:size(A,1)
vec(:,:,iter) = reshape(A(iter,2:end), [n, n]);
end
% loop over each row
for k = 1:size(A,1)
vec_ = vec(:,:,k);
temp = [];
[row, col, val] = find(vec_); % Use find to get the non-zero values and their indices
% Replace -1 with 2
val(val == -1) = 2;
temp = [col, row, val]; % Concatenate the columns and values
% Add your code for write to file and create a separate directory
end
