Binmat: On-disk matrix class using on-demand disk access

Description

The Binmat class implements on-disk matrices with efficient access to columns. Values within each column are contiguously stored in a binary file on disk. The columns themselves need not be stored contiguously. Only the accessed elements of the matrix are loaded into memory.

New code should use the matter_mat class from the matter package instead.

Usage

## Instance creation
Binmat(
    files,
    nrow, ncol,
    offsets = 0,
    extents = rep(nrow, ncol),
    datatype = c("16-bit integer",
        "32-bit integer",
        "64-bit integer",
        "32-bit float",
        "64-bit float"),
    dimnames = NULL,
    ...)

## Additional methods documented below

Arguments

Slots

files:

A factor giving the full file paths of the binary files storing the matrix (or matrices) on disk. Length must be equal to the number of columns.

offsets:

A numeric vector giving the positions of the first value of each column in number of bytes from the beginning of the file.

extents:

A numeric vector giving the length of each column.

datatype:

A factor vector giving the binary data types of each element of the matrix (or matrices) on disk. Length must be equal to the number of columns.

dim:

A length 2 integer vector analogous to the 'dim' attribute of an ordinary R matrix.

dimnames:

A length 2 list analogous to the 'dimnames' attribute of an ordinary R matrix.

.__classVersion__:

A Versions object describing the version of the class used to created the instance. Intended for developer use.

Extends

Versioned

Creating Objects

Binmat instances are usually created through Binmat().

Methods

Standard generic methods:

dim(x), dim(x) <- value:

Return or set the dimensions of the on-disk matrix.

dimnames(x), dimnames(x) <- value:

Return or set the 'dimnames' of the on-disk matrix.

colnames(x), colnames(x) <- value:

Return or set the column names of the on-disk matrix.

rownames(x), rownames(x) <- value:

Return or set the row names of the on-disk matrix.

ncol:

Return the number of columns in the on-disk matrix.

nrow:

Return the number of columns in the on-disk matrix.

cbind:

Combine on-disk matrices by columns.

rbind:

Not allowed for on-disk matrices. (Always returns an error.)

Binmat[i, j, ..., drop]:

Access elements in the on-disk matrix. A Binmat on-disk matrix can be indexed like an ordinary R matrix. Note however that linear indexing is not supported. Assignment is not currently allowed.

Author(s)

Kylie A. Bemis

See Also

matrix, Hashmat, SImageSet

Examples

## Not run: 
## Create an Binmat object
Binmat("path/to/file.extension")

## End(Not run)