rechunkByMargins: Rechunk by margins

Description Usage Arguments Details Value Author(s) Examples

View source: R/rechunkByMargins.R

Description

Convert an existing HDF5Matrix into a pure column- or row-based chunk layout.

Usage

1
2
rechunkByMargins(x, size=5000, outfile=NULL, outname=NULL, 
    outlevel=NULL, byrow=TRUE) 

Arguments

x

A HDF5Matrix object.

size

An integer scalar specifying the number of elements in each chunk.

outfile

A string containing the name for the output HDF5 file, chosen by getHDF5DumpFile if not specified.

outname

A string containing the name for the output HDF5 data set, chosen by getHDF5DumpName if not specified.

outlevel

An integer scalar specifying the compression level, chosen by getHDF5DumpCompressionLevel if not specified.

byrow

A logical scalar indicating if the output file should be row-chunked (default) or column-chunked.

Details

Pure column- or row-based chunk layouts are optimal for random column and row access, respectively, from a HDF5 file. This function can be used to convert a file into a pure row/column layout prior to calling other functions. In many cases, a small investment in rechunking time is repaid by a reduction in access times in downstream procedures.

Value

A HDF5Matrix object pointing to the HDF5 file containing the data from x but with the new chunk layout.

Author(s)

Aaron Lun

Examples

1
2
3
A <- as(matrix(runif(5000), nrow=100, ncol=50), "HDF5Array")
byrow <- rechunkByMargins(A, byrow=TRUE)
bycol <- rechunkByMargins(A, byrow=FALSE)

beachmat documentation built on Nov. 1, 2018, 4:22 a.m.