/ 
Statistics-R-IO-0.07
River stage zero No dependents
/ Statistics::R::IO

NAME

Statistics::R::IO - Perl interface to serialized R data

VERSION

version 0.07

SYNOPSIS

use Statistics::R::IO;

my $var = Statistics::R::IO::readRDS('file.rds');
print $var->to_pl;

my %r_workspace = Statistics::R::IO::readRData('.RData');
while (my ($var_name, $value) = each %r_workspace) {
    print $var_name, $value;
}

my $pi = Statistics::R::IO::evalRserve('pi');
print $pi->to_pl;

DESCRIPTION

This module is a pure-Perl implementation for reading native data files produced by the R statistical computing environmnent)

It provides routines for reading files in the two primary file formats used in R for serializing native objects:

RDS

RDS files store a serialization of a single R object (and, if the object contains references to other objects, such as environments, all the referenced objects as well). These files are created in R using the readRDS function and are typically named with the .rds file extension.

RData

RData files store a serialization of a collection of named objects, typically a workspace. These files are created in R using the save function and are typically named with the .RData file extension. (Contents of the R workspace can also be saved automatically on exit to the file named .RData, which is by default automatically read in on startup.)

As of version 0.04, the module can also evaluate R code on a remote host that runs the Rserve binary R server. This allows Perl programs to access all facilities of R without the need to have a local install of R or link to an R library.

See "SUBROUTINES" for invocation and usage information on individual subroutines, and the R Internals manual for the specification of the file formats.

EXPORT

Nothing by default. Optionally, subroutines readRDS, readRData, and evalRserve, or :all for all three.

SUBROUTINES

readRDS EXPR

Reads a file in RDS format whose filename is given by EXPR and returns a Statistics::R::REXP object.

readRData EXPR

Reads a file in RData format whose filename is given by EXPR and returns a hash whose keys are the names of objects stored in the file with corresponding values as Statistics::R::REXP instances.

evalRserve REXPR [ HOSTNAME [, PORT] | HANDLE]

Evaluates an R expression, given as text string in REXPR, on an Rserve server and returns its result as a Statistics::R::REXP object.

The server location can be specified either by its host name and (optionally) port or by a connected instance of IO::Handle. The caller passing the HANDLE is responsible for reading (and checking) the server ID that is returned in the first 32-byte response when the connection was established. This allows opening the connection once and reusing it in multiple calls to 'evalRserve'.

If only REXPR is given, the function assumes that the server runs on the localhost. If PORT is not specified, it defaults to the standard Rserve port, 6311.

The function will close the connection to the Rserve host if it has opened it itself, but not if the connection was passed as a HANDLE.

DEPENDENCIES

Requires perl 5.012 or newer.

Core modules

  • strict

  • warnings

  • overload

  • Carp

  • Exporter

  • Module::Build

  • Scalar::Util

  • Test::More

Additional CPAN modules

  • Moo

  • namespace::clean

  • Test::Fatal

BUGS AND LIMITATIONS

The module currently handles the 'version 2' serialization format, used since R 1.4.0 (released in December 2001). Only XDR and native-order binary is implemented, and since the R documentation describes the ASCII save format as "now mainly of historical interest", this is unlikely to change soon. No check is performed that a file stored in native-order binary was created on a platform that used the same order, and it is up to the caller to ensure compatibility. (Given that the default save format is XDR, and the prevalence of Intel platforms, this is unlikely to be a problem for either publicly-distributed or internal data files.)

Data files compressed with 'gzip' and 'bzip2' are supported, but not 'xz' ones. Again, given the R defaults ('gzip') and the fact that IO::Uncompress::UnXz is not production-ready, this is unlikely to change soon.

There are some R types that are not (yet) implemented, although all typical "user-facing" types -- such as vectors, lists, and environments -- are. The remaining R types will be implemented as-needed; in other words, if you come across one that you need to read a particular file, please send me the type (the id will included in the "unimplemented SEXPTYPE" error message) and, if possible, how it was generated.

There are no known bugs in this module. Please report any bugs or feature requests to bug-statistics-r-io at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Statistics-R-IO. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

SUPPORT

You can find documentation for this module with the perldoc command.

perldoc Statistics::R::IO

You can also look for information at:

AUTHOR

Davor Cubranic <cubranic@stat.ubc.ca>

COPYRIGHT AND LICENSE

This software is Copyright (c) 2014 by University of British Columbia.

This is free software, licensed under:

The GNU General Public License, Version 3, June 2007