perlcpanstrawberry-perlmetacpan

To Convert CSV File to Hash Structure using TEXT::CSV_XS Module in perl


I've below code which is used to read a csv file and convert to hash. The Keys are dependent on the number of key columns that user need.

use warnings;
use strict;

my %hash;     
my $KeyCols = 2;
while (<DATA>) {
    chomp;
    my @cols = split /,/, $_, $KeyCols+1;
    next unless @cols > $KeyCols;
    my $v = pop @cols;
    my $k = join '', @cols;
    $hash{$k} = $v;
}

I need help in achieving the same logic using TEXT::CSV_XS package for efficiency. Please help.


Solution

  • The real reason for using Text::CSV_XS is for correctness. It's not going to be faster than what you have, but it will work where yours will fail.

    use Text::CSV_XS qw( );
    
    my $csv = Text::CSV_XS->new({
        auto_diag => 2,
        binary    => 1,
    });
    
    my %hash;
    while ( my $row = $csv->getline(\*DATA) ) { 
       $hash{ $row->[0] . $row->[1] } = $row;
    }
    

    Concatenating the fields together directly (without a separator) seems really odd.

    The above makes the value an array of fields rather than CSV. If you want CSV as in the original, you will need to re-encode them into CSV.

    my %hash;
    while ( my $row = $csv->getline(\*DATA) ) { 
       my ($k1, $k2) = splice(@$row, 0, 2);
       $csv->combine(@$row);
       $hash{ $k1 . $k2 } = $csv->string();
    }