Making the result of `nullspace` sparse

Is it possible to make the result of LinearAlgebra.nullspace as sparse as possible? How about when using the package Nemo for sparse matrices over \mathbb{R}?

Regardless of whether or not this is possible, what would be the theory behind this? From the documentation of LinearAlgebra.nullspace, it’s based on the SVD of the matrix, but I’m not sure where the sparsity would come in, if that’s even possible.


I just wanted to point out, that Nemo unfortunately does not have support for sparse matrices.