I have two data files. The data.txt file is tab-delimited and does not contain blank entries.

1    6    11
2    7    12
3    8    13
4    9    14
5    10    15
[download]

The data_missing.txt file is tab-delimited but contains some blank entries.

1    6    11
2    7    
3    8    13
4        14
5    10    15
[download]

I use the following script to test the contents of the pdls created by rcols:

#!/usr/bin/env perl

use strict;
use warnings;

use PDL;
use PDL::IO::Misc;

my $file_name = shift;

die 'No file given.' unless defined($file_name);
open(my $fh, '<', $file_name) or die "Can not open file: $!";

my @pdls = rcols $fh, { COLSEP => "\t" };

foreach (@pdls) {
    print "$_\n";
}

exit;
[download]

The output for <data.txt> is:

[1 2 3 4 5]
[6 7 8 9 10]
[11 12 13 14 15]
[download]

The output for <data_missing.txt> is:

[1 2 3 4 5]
[6 7 8 0 10]
[11 0 13 14 15]
[download]

So far, so good. However, if I change the value for $PDL::undefval, I get a strange result. First, the default value of $PDL::undefval is zero.

perl -MPDL -E 'say $PDL::undefval'
0
[download]

Here is the code with $PDL::undefval set to -999.

#!/usr/bin/env perl

use strict;
use warnings;

use PDL;
use PDL::IO::Misc;

my $file_name = shift;

die 'No file given.' unless defined($file_name);
open(my $fh, '<', $file_name) or die "Can not open file: $!";

local $PDL::undefval = -999;

my @pdls = rcols $fh, { COLSEP => "\t" };

foreach (@pdls) {
    print "$_\n";
}

exit;
[download]

The output for <data.txt> is:

[1 2 3 4 5]
[6 7 8 9 10]
[11 12 13 14 15]
[download]

The output for <data_missing.txt> is:

[1 2 3 4 5]
[6 7 8 0 10]
[11 -999 13 14 15]
[download]

The value of $PDL::undefval is used in one case (where the 12 was deleted at the end of a row in the input file), but a zero used (where the 9 was deleted in the middle of a row in the input file).