Eigen Sparse valuePtr is displaing zeros, while leaving out valid values - c++

Question

I don't understand the result I get when I try to iterate over valuePtr of a sparse matrix. Here is my code.
#include <iostream>
#include <vector>
#include <Eigen/Sparse>
using namespace Eigen;
int main()
{
SparseMatrix<double> sm(4,5);
std::vector<int> cols = {0,1,4,0,4,0,4};
std::vector<int> rows = {0,0,0,2,2,3,3};
std::vector<double> values = {0.2,0.4,0.6,0.3,0.7,0.9,0.2};
for(int i=0; i < cols.size(); i++)
sm.insert(rows[i], cols[i]) = values[i];
std::cout << sm << std::endl;
int nz = sm.nonZeros();
std::cout << "non_zeros : " << nz << std::endl;
for (auto it = sm.valuePtr(); it != sm.valuePtr() + nz; ++it)
std::cout << *it << std::endl;
return 0;
}
Output:
0.2 0.4 0 0 0.6 // The values are in the matrix
0 0 0 0 0
0.3 0 0 0 0.7
0.9 0 0 0 0.2
non_zeros : 7
0.2 // but valuePtr() does not point to them
0.3 // I expected: 0.2, 0.3, 0.9, 0.4, 0.6, 0.7, 0.2
0.9
0
0.4
0
0
I don't understand why I am getting zeros, what's going on here?

Answer 1

According to the documentation for SparseMatrix:
Unlike the compressed format, there might be extra space inbetween the
nonzeros of two successive columns (resp. rows) such that insertion of
new non-zero can be done with limited memory reallocation and copies.
[...]
A call to the function makeCompressed() turns the matrix into the standard compressed format compatible with many library.
For example:
This storage scheme is better explained on an example. The following
matrix
0 3 0 0 0
22 0 0 0 17
7 5 0 1 0
0 0 0 0 0
0 0 14 0 8
and one of its possible sparse, column major representation:
Values: 22 7 _ 3 5 14 _ _ 1 _ 17 8
InnerIndices: 1 2 _ 0 2 4 _ _ 2 _ 1 4
[...]
The "_" indicates available free space to quickly insert new elements.
Since valuePtr() simply return a pointer to the Values array, you'll see the empty spaces (the zeroes that got printed) unless you make the matrix compressed.