README.md

In stela2502/Dense2SparseHDF5: Memory efficient convert a dens HDF5 matrix to a sparse matrix

Dense2SparseHDF5

This R package is meant to easily convert a HDF5 dense matrix, which is found in several .loom files and can easily be too big for a normal analysis desktop (64Gb RAM), into a R sparse matrix.

The process in implemented in c++ using the Rcpp package, but needs further speed improvements.

Install

install.packagaes('devtools')

devtools::install_git('https://github.com/stela2502/Dense2SparseHDF5.git')

Usage

library(Dense2SparseHDF5)

file ="someFile.loom"

obj= Dense2SparseHDF5$new(file)

obj$toSparseVector( slot='matrix', block_size=1000 )

obj$Matrix

If the default approach fails try to change the block_size from the default 1e+3 to some other value (e.g. 1e+4).

stela2502/Dense2SparseHDF5 documentation built on March 19, 2020, 1:08 a.m.