How to recreate sparse matrix from vector of CsVec efficiently?

[newtonharry]

Hi,

I'm currently recreating a sparse matrix from a vector of CsVec's and I'm wondering if I'm taking the optimal approach here:

use sprs::{CsMat, CsVec};

fn main() {
    let sparse_cols: Vec<CsVec<i64>> = /* ... */;
    let rows = /* number of rows in the original matrix */;

    println!("Converting back to a sparse csc matrix");
    let mut data = Vec::new();
    let mut indices = Vec::new();
    let mut indptr = vec![0]; 

    for col in sparse_cols.iter() {
        let col_indices = col.indices();
        let col_data = col.data();
        data.extend_from_slice(col_data);
        indices.extend_from_slice(col_indices);
        indptr.push(indptr.last().unwrap() + col.nnz() as i64);
    }

    let merged_matrix = CsMat::new_csc((matrix.rows(), sparse_cols.len()), indptr, indices, data);
    println!("Merged matrix shape {:?}", merged_matrix.shape());
}

Thanks!

[mulimoen]

This is pretty much the same as done by repeated calls to append_outer_csvec. You can make it slightly more efficient by pre-allocating using Vec::with_capacity