The documentation for the nonZeros() method on a dense matrix states:

the number of nonzero coefficients which is in practice the number of stored coefficients.

The following code outputs 10, however:

Code: Select all

std::cout << VectorXs::Zero(10).nonZeros() << std::endl;

I take it for a dense matrix nonZeros() always gives the size of the matrix, despite the documentation stating that it is the "the number of nonzero coefficients"? Not a big deal, I can do something like this instead, but it seems like it could be inefficient:

Code: Select all

( matrix.array() == 0 ).cast<int>().sum()

Thanks for the clarification!

Right, the documentation is not very clear as this method is useful for sparse matrices only, and available on dense matrices for API compatibility. Here "non-zeros" really means explicitly stored coefficients. You can use the count() method to simplify the expression:

Code: Select all

(matrix.array() == 0).count()