compare the value of same items with one exception case - perl

Question

A,food,75
B,car,136
A,car,69
B,shop,80
A,house,179
B,food,75
C,car,136
ECX5,flight,50
QC4,train,95
C,food,85
B,house,150
D,shop,80
EAX5,flight,50
QA4,train,75
F,movie,
It should do comparison between the values of same type (wherever the 2nd column matches) and Print the differ .Now i want output to look like:
**A,food,75 is not matching with B,food,75 C,food,85
A,car,69 is not matching with C,car,136 B,Car,136
A,house,179 is not matching with B,house,150
QC4,train,95 is not matching with QA4,train,75
F,movie missing value
Code I've written is below but its not printing the format the way I want.
while (FILE) {
my $line = $_ ;
my #lines = split /,/, $line ;
$data{$lines[1]}{$lines[0]} = $lines[2] ;
}
foreach my $item (keys %val) {
foreach my $letter1 (keys %{$val{$item}}) {
foreach my $letter2 (keys %{$val{$item}}) {
if ( ($val{$item}{$letter1} != $val{$item}{$letter2}) && ($letter1 ne
$letter2) && ( (!defined $done{$item}{$letter1}{$letter2}) ||
(!defined
$done{$item}{$letter2}{$letter1}) ) ) {
print "$item : $letter1, $val{$item}{$letter1}, $letter2 ,
$val{$item}
{$letter2}\n" ;
}
}

Answer 1

Really hard to follow the logic of your code.
But I seem to get the desired result with this:
[Edit] The code was edited as per the comments
use strict;
use warnings;
my (%hash,
);
while(my $line=<DATA>) {
chomp $line;
my ($letter, $object, $number)=split /,/, $line;
### here we are dealing with missing values
if ($number =~ m{^\s*$}) {
print $line, "missing value\n";
next;
}
### here we dissever exceptional items apart from the others
if ($letter=~m{^E[AC]X\d$}) {
$object = "exceptional_$object";
}
$number+=0; # in case there is whitespace at the end
push #{$hash{$object}{$number}}, [$letter,$number,$line];
}
for my $object(sort keys %hash) {
my $oref = $hash{$object};
if (1==keys %$oref) {
next;
}
my $str;
for my $item (values %$oref) {
$str .= $str ? " $item->[0][2]" : "$item->[0][2] is not matching with";
}
print ($str,"\n");
}
__DATA__
A,food,75
B,car,136
A,car,69
B,shop,80
A,house,179
B,food,75
C,car,136
ECX5,flight,50
ECX4,train,95
C,food,85
B,house,150
D,shop,80
EAX5,flight,50
EAX4,train,75
F,movie,
output
F,movie,missing value
A,car,69 is not matching with B,car,136
EAX4,train,75 is not matching with ECX4,train,95
C,food,85 is not matching with A,food,75
A,house,179 is not matching with B,house,150
What the algorithm does:
Looping through the input we remember the all the lines for each unique pair of object and number.
After going through the input loop we do the following:
For each object we skip it if it has no different numbers:
if (1==keys %$oref) {
next
}
If it has, we build an output string from a list of the first remembered lines for that object and number (that is we omit the duplicates for the object and number);
the first item from the list amended with "is not matching with".
Also, I am reading from the special filehandle DATA, which accesses embedded data in the script. This is for convenience of demoing the code