Proof of concept to show the efficiency of using the sparse representation.
As commented in the source code, this uses a dgCMatrix, a compressed, sparse, column-oriented numeric (double) matrix from the Matrix package. What we really want to use to store these data is a igCMatrix, a compressed, sparse, column-oriented integer matrix. However, the igCMatrix class, while defined in the Matrix package, is not actually implemented. Matrix is in DEPENDS because you can't do much with the returned object otherwise (I think).
These changes caused some havoc with roxygen, so I needed to do some manual editing. I'll try to figure out the source and fix so this doesn't have to be done each time.
Also, where can I obtain v2.19.3-2 of rhdf5? Can't see it on Github.
Proof of concept to show the efficiency of using the sparse representation.
As commented in the source code, this uses a dgCMatrix, a compressed, sparse, column-oriented numeric (double) matrix from the Matrix package. What we really want to use to store these data is a igCMatrix, a compressed, sparse, column-oriented integer matrix. However, the igCMatrix class, while defined in the Matrix package, is not actually implemented. Matrix is in
DEPENDS
because you can't do much with the returned object otherwise (I think).These changes caused some havoc with roxygen, so I needed to do some manual editing. I'll try to figure out the source and fix so this doesn't have to be done each time.
Also, where can I obtain
v2.19.3-2
of rhdf5? Can't see it on Github.