scalar(@$colref) is the number of items found in the column, it would never contain a '#' symbol.

$colref->[0] is the first item on the line read in, it could contain a # symbol.

if ($colref->[0] =~ /^#/) { next; }
[download]

Notice i said next rather than last. next will go on to the next row, last will end the do loop and stop reading any more lines.

the next condition is a touch more tricky.

 
my $anynonblank=0; 
for my $item (@$colref){ 
  unless  ($item =~ /^\s+$/) { $anynonblank=1; last; } 
} 
unless ($anynonblank) { next; }
[download]

edit: opps , fixed as per Re^6: Read CSV with column mapping below

unless ($item =~ /^s+$/) { ... }

The  $item =~ /^s+$/ expression matches against a  's' literal character and not, as I think was intended, the  \s whitespace character class.

---

Give a man a fish:  <%-{-{-{-<

I have tried your suggestion and it did help to avoid comments but not for blank lines. Here I am provide the code and config file. Also providing output.

use strict;
use warnings;
use Data::Dumper;
use Text::CSV_XS;
use IO::File;

my $hash_ref = csv_file_hashref('parameters.conf');

foreach my $key (sort keys %{$hash_ref}){
   print qq{$key:};
   print join q{,}, @{$hash_ref->{$key}};
   print qq{\n};
}

sub csv_file_hashref {
   my ($filename) = @_;

   my $csv_fh = IO::File->new($filename, 'r');
   my $csv = Text::CSV_XS->new ();

   my %output_hash;
   while(my $colref = $csv->getline ($csv_fh))
   {
      if ($colref->[0] =~ /^#/) {
        next;
        }
   my $anynonblank=0; 
    for my $item (@$colref){ 
        unless  ($item =~ /^\s+$/) { $anynonblank=1; last; } 
    } 
    unless ($anynonblank) { next; }  
      #print @$colref;
      #print "\n";
      $output_hash{shift @{$colref}} = $colref;
   }

   return \%output_hash;
}
1;
[download]

config file-

#sdadwlasda
abc xyz,10

#ldfgld

xyz pqr,2

#jwej

pqr stq,0.1

#
#

#
#
[download]

Output-

:
abc xyz:10
pqr stq:0.1
xyz pqr:2
[download]

Two alternatives:

use 5.14.1;
use warnings;

use Data::Peek;
use Text::CSV_XS "csv";

my %hash;
csv (in => "pm1227312.csv", out => undef, on_in => sub {
    my ($key, $value) = @{$_[1]};
    $key && $key !~ m/^\s*(?:#|$)/ and $hash{$key} = $value;
    });
DDumper \%hash;

%hash = map { @$_ } @{csv (in => "pm1227312.csv", filter => { 0 => sub
+ { !m/^\s*(?:#|$)/ }})};
DDumper \%hash;
[download]


Enjoy, Have FUN! H.Merijn

/^\s+$/ requires at least one blank character. /^\s*$/ would fail empty lines too.

You could just check if the key is blank

#!/usr/bin/perl
use strict;
use warnings;
use Data::Dumper;
use Text::CSV_XS;
use IO::File;

my $hash_ref = csv_file_hashref('parameters.conf');

foreach my $key (sort keys %$hash_ref){
  print $key.':';
  print join ',', @{$hash_ref->{$key}};
  print "\n";
}

sub csv_file_hashref {
  my ($filename) = @_;
  my %output_hash = ();

  my $csv_fh = IO::File->new($filename, 'r');
  my $csv = Text::CSV_XS->new( { allow_whitespace => 1 } );

  while (my $row = $csv->getline($csv_fh)){
    my $key = shift @$row;
    next if $key =~ /^#/;
    if ($key =~ /\S/){
      $output_hash{$key} = $row ;
    }
  }

  return \%output_hash;
}
[download]

Note $item =~ /^\s+$/ matches 1 or more white space only (but not nothing). Try $item =~ /^\s*$/ in your code.

poj