| Subject: | Convenient utilities |
Hi,
I have written some code that makes my life with Statistics::R easier
and I thought I'd share (attached).
First, I added a run() method that does the following:
i/ use send() to send a command to R
ii/ use read() to read the output of R
iii/ checks that R processed the command successfully or die()
iv/ return the R output
The other method is value(), which, given an R variable name queries R
and return its value (without decoration, column names, ...). This works
for scalars, and an obvious extension would be to support R lists and to
return Perl arrays.
Any chance these methods can be officially added to Statistics::R? Thanks,
Florent
| Subject: | R.pm |
package Statistics::R;
sub run {
my ($self, $cmd) = @_;
# Execute command
$self->send($cmd);
# Read command output
my $output = $self->read;
# Check for errors
if ($output =~ m/^<.*>$/) {
die "Error: There was a problem running the R command\n".
" $cmd\n".
"Because\n".
" $output\n";
}
return $output;
}
sub value {
# Get the value of a variable through a Statistics::R object
my ($self, $varname) = @_;
$self->send(qq`print($varname)`);
my $string = $self->read;
my $value;
if ($string eq 'NULL') {
$value = undef;
} elsif ($string =~ m/^\s*\[\d+\]/) {
# String look like: ' [1] 6.4 13.3 4.1 1.3 14.1 10.6 9.9 9.6 15.3
# [16] 5.2 10.9 14.4'
my @lines = split /\n/, $string;
for (my $i = 0; $i < scalar @lines; $i++) {
$lines[$i] =~ s/^\s*\[\d+\] //;
}
$value = join ' ', @lines;
# may need to split array and remove quotes
$value =~ s/^"(.*)"$/$1/;
} else {
my @lines = split /\n/, $string;
if (scalar @lines == 2) {
# String looks like: ' mean
# 10.41111 '
# Extract value from second line
$value = $lines[1];
$value =~ s/^\s*(\S+)\s*$/$1/;
} else {
die "Error: Don't know how to handle this R output\n$string\n";
}
}
return $value;
}
1;