Eigen SparseMatrix upper triangular to full matrix

Question

I have an upper triangular SparseMatrix<double>. What would be the most efficient way to convert it to a full sparse matrix?

I have this implemented currently as mat.transpose() + mat - diagonal(mat).

I thought I could use something like

mat.selfadjointView<Eigen::Lower>() = mat.selfadjointView<Eigen::Upper>();

For reasons I don't fully understand, this clears the matrix.

Answer 1

According to the documentation for Eigen::MatrixBase::selfadjointview, the function already creates a symmetric view from the upper or lower triangular part.

Matrix3i m = Matrix3i::Random();
cout << "Here is the matrix m:" << endl << m << endl;
cout << "Here is the symmetric matrix extracted from the upper part of m:" << endl
     << Matrix3i(m.selfadjointView<Upper>()) << endl;
cout << "Here is the symmetric matrix extracted from the lower part of m:" << endl
     << Matrix3i(m.selfadjointView<Lower>()) << endl;

Output:

Here is the matrix m:
 7  6 -3
-2  9  6
 6 -6 -5
Here is the symmetric matrix extracted from the upper part of m:
 7  6 -3
 6  9  6
-3  6 -5
Here is the symmetric matrix extracted from the lower part of m:
 7 -2  6
-2  9 -6
 6 -6 -5

Assuming your matrix is upper triangular, the following should answer your question.

Matrix3i m = [] {
   Matrix3i tmp;
   tmp << 1, 2, 3, 0, 4, 5, 0, 0, 6;
   return tmp;
}();
cout << "Here is the matrix m:" << endl << m << endl;
cout << "Here is the symmetric matrix extracted from the upper part of m:" << endl
     << Matrix3i(m.selfadjointView<Upper>()) << endl;

Output:

Here is the matrix m:
 1  2  3
 0  4  5
 0  0  6
Here is the symmetric matrix extracted from the upper part of m:
 1  2  3
 2  4  5
 3  5  6