Re: Matching probem

[kcott]

G'day Nansh,

I'd consider using a hash. Use the makes for the main keys; you can always generate an array of them with keys if needed.

You didn't explain what the lines starting with "*." were supposed to be. I've assumed some sort of comment and labelled them as such.

You didn't explain why some models start with a "+": I've put them in separate arrays.

#!/usr/bin/env perl

use strict;
use warnings;

use Data::Dump;

my %cars;
my $models_re = qr{(?x: ^ ( \w+ ) \s+ ( [^*]+? ) \s+ [*][.] ( .*? ) \s
+* $ )};
my $plus_re = qr{(?x: ^ [+] ( .* ) $ )};

{
    local $/ = 'CARS ';

    while (<DATA>) {
        chomp;
        next unless length;
        my ($make, $models, $comment) = /$models_re/;
        
        for (split /\s+/, $models) {
            if (/$plus_re/) {
                push @{$cars{$make}{models}{plus}}, $1;
            }
            else {
                push @{$cars{$make}{models}{plain}}, $_;
            }
        }

        $cars{$make}{comment} = $comment;
    }
} 

dd \%cars;

__DATA__
CARS TOYOTO Etios Innova Camry

+Fortuner Corolla

*.This is Class cars.

CARS HYUNDAI Verna i10 i20

+Eon Xcent

+Tucson

*.It is best cars.

CARS HONDA City Brio Amaze

+Accord

*.This is Luxury cars.

Output:

{
  HONDA   => {
               comment => "This is Luxury cars.",
               models  => { plain => ["City", "Brio", "Amaze"], plus =
+> ["Accord"] },
             },
  HYUNDAI => {
               comment => "It is best cars.",
               models  => {
                            plain => ["Verna", "i10", "i20", "Xcent"],
                            plus  => ["Eon", "Tucson"],
                          },
             },
  TOYOTO  => {
               comment => "This is Class cars.",
               models  => {
                            plain => ["Etios", "Innova", "Camry", "Cor
+olla"],
                            plus  => ["Fortuner"],
                          },
             },
}