realize: Realize an object in memory or on disk

In Bioconductor/DelayedArray: A unified framework for working transparently with on-disk and in-memory array-like datasets

Realize an object in memory or on disk

Description

realize() is an S4 generic function.

The default realize() method handles the array case. It will realize the array-like object (typically a DelayedArray object) in memory or on disk, depending on the realization backend specified via its BACKEND argument,

Usage

realize(x, ...)

## S4 method for signature 'ANY'
realize(x, BACKEND=getAutoRealizationBackend())

Arguments

x	An array-like object (typically a DelayedArray object) for the default method. Other types of objects can be supported via additional methods. For example, the SummarizedExperiment package defines a method for SummarizedExperiment objects (see ?`realize,SummarizedExperiment-method`).
...	Additional arguments passed to methods.
BACKEND	NULL or a single string specifying the name of a realization backend. By default, the automatic realization backend will be used. This is the backend returned by getAutoRealizationBackend().

Details

The default realize() method realizes an array-like object x in memory if BACKEND is NULL, otherwise on disk.

Note that, when BACKEND is not NULL, x gets realized as a "pristine" DelayedArray object (e.g. an HDF5Array object), that is, as a DelayedArray object that carries no delayed operations. This means that, if x is itself a DelayedArray object, then the returned object is another DelayedArray object semantically equivalent to x where the delayed operations carried by x have been realized.

Value

A "pristine" DelayedArray object if BACKEND is not NULL.

Otherwise, an ordinary matrix or array, or a SparseArraySeed object.

See Also

• getAutoRealizationBackend and setAutoRealizationBackend for getting and setting the automatic realization backend.

• DelayedArray objects.

• RleArray objects.

• HDF5Array objects in the HDF5Array package.

• array objects in base R.

Examples

## ---------------------------------------------------------------------
## In-memory realization
## ---------------------------------------------------------------------

a <- array(1:24, dim=4:2)
realize(a, BACKEND=NULL)  # no-op

A <- DelayedArray(a)
realize(log(A), BACKEND=NULL)  # same as 'as.array(log(A))'

## Sanity checks:
stopifnot(identical(realize(a, BACKEND=NULL), a))
stopifnot(identical(realize(log(A), BACKEND=NULL), log(a)))

## ---------------------------------------------------------------------
## On-disk realization
## ---------------------------------------------------------------------

library(HDF5Array)

realize(log(A), BACKEND="HDF5Array")  # same as 'as(log(A), "HDF5Array")'

## ---------------------------------------------------------------------
## Omitting the 'BACKEND' argument
## ---------------------------------------------------------------------

## When 'BACKEND' is not specified, the "automatic realization backend"
## is used. This backend is controlled via setAutoRealizationBackend().

toy_h5 <- system.file("extdata", "toy.h5", package="HDF5Array")
h5ls(toy_h5)
M1 <- HDF5Array(toy_h5, "M1")
M2 <- HDF5Array(toy_h5, "M2")
M3 <- rbind(log(M1), t(M2)) + 0.5
M3

## Set the "automatic realization backend" to NULL for in-memory
## realization (as ordinary array or SparseArraySeed object):
setAutoRealizationBackend(NULL)
m3 <- realize(M3)  # in-memory realization

registeredRealizationBackends()

setAutoRealizationBackend("RleArray")
realize(M3)  # realization as RleArray object

setAutoRealizationBackend("HDF5Array")
realize(M3)  # on-disk realization (as HDF5Array object)

setAutoRealizationBackend()  # restore default (NULL)

Bioconductor/DelayedArray documentation built on March 4, 2024, 9:12 p.m.