bigpca (version 1.1)

big.select: Select a subset of a big.matrix

Description

Select a subset of big.matrix using indexes for a subset of rows and columns. Essentially a wrapper for bigmemory::deepcopy, but with slightly more flexible parameters. bigMat can be entered in any form accepted by get.big.matrix(), row and column selections can be vectors of indexes, names or file.names containing indexes. Default is to process using deepcopy, but processing without using bigmemory native methods is a faster option when matrices are small versus available RAM. File names for backing files are managed only requiring you to enter a prefix, or optionally use the default and gain filebacked functionality without having to bother choosing filename parameters.

Usage

big.select(bigMat, select.rows = NULL, select.cols = NULL, dir = getwd(),
  deepC = TRUE, pref = "sel", delete.existing = FALSE, verbose = FALSE)

Arguments

bigMat

a big.matrix, matrix or any object accepted by get.big.matrix()

select.rows

selection of rows of bigMat, can be numbers, logical, rownames, or a file with names. If using a filename argument, must also use a filename argument for select.cols (cannot mix)

select.cols

selection of columns of bigMat, can be numbers, logical, colnames, or a file with names

dir

the directory containing the bigMat backing file (e.g, parameter for get.big.matrix()).

deepC

logical, whether to use bigmemory::deepcopy, which is slowish, but scalable, or alternatively to use standard indexing which converts the result to a regular matrix object, and is fast, but only feasible for matrices small enough to fit in memory.

pref

character, prefix for the big.matrix backingfile and descriptorfile, and optionally an R binary file containing a big.matrix.descriptor object pointing to the big.matrix result.

delete.existing

logical, if a big.matrix already exists with the same name as implied by the current 'pref' and 'dir' arguments, then default behaviour (FALSE) is to return an error. to overwrite any existing big.matrix file(s) of the same name(s), set this parameter to TRUE.

verbose

whether to display extra information about processing and progress

Value

A big.matrix with the selected (in order) rows and columns specified

Examples

# NOT RUN {
orig.dir <- getwd(); setwd(tempdir()); # move to temporary dir
if(file.exists("sel.bck")) { unlink(c("sel.bck","sel.dsc")) }
bmat <- generate.test.matrix(5,big.matrix=TRUE)
# take a subset of the big.matrix without using deepcopy
sel <- big.select(bmat,c(1,2,8),c(2:10),
 deepC=FALSE,verbose=TRUE, delete.existing=TRUE)
prv.big.matrix(sel)
# now select the same subset using row/column names from text files
writeLines(rownames(bmat)[c(1,2,8)],con="bigrowstemp.txt")
writeLines(colnames(bmat)[c(2:10)],con="bigcolstemp.txt")
sel <- big.select(bmat, "bigrowstemp.txt","bigcolstemp.txt", delete.existing=TRUE, pref="sel2")
prv.big.matrix(sel)
rm(bmat)
rm(sel)  
unlink(c("bigcolstemp.txt","bigrowstemp.txt","sel.RData","sel2.bck","sel2.dsc"))
setwd(orig.dir) # reset working dir to original
# }