Perlでの処理について質問です．

このフォーマットであればtombeさんのやり方で十分な気もしますが，CSVファイルの解析として取り組むならこんな感じでしょうか。

#! /usr/bin/perl

package	Object;

use	strict;
use	base qw(Class::Accessor);

sub new {
	my	($pkg) = @_;
	bless({},$pkg);
}


package	Company;

use	strict;
use	base qw(Object);

__PACKAGE__->mk_accessors(qw(name data parsed));

sub initialize {
	my	($self) = @_;
	$self->data([]);
	$self->parsed([]);
	$self;
}

sub parse {
	my	($self) = @_;
	my	$i = 0;
	my	@parsed = ();
	foreach my $data (@{$self->data()}) {
		if ($data !~ m/^a/) {
			$i++;
		} else {
			push(@parsed,$i);
			$i = 0;
		}
	}
	$self->parsed(\@parsed);
	$self;
}


package	CompanyIO;

use	strict;
use	Text::CSV;
use	IO::File;
use	base qw(Object);

__PACKAGE__->mk_accessors(qw(type file_name file csv));

sub initialize {
	my	($self) = @_;
	my	$file = new IO::File();
	my	$csv = new Text::CSV({'binary'=>1});
	$file->open($self->file_name(),$self->type()) || die "cannot open file";
	$self->file($file);
	$self->csv($csv);
	$self;
}

sub finalize {
	my	($self) = @_;
	$self->file()->close();
}


package	CompanyReader;

use	strict;
use	Text::CSV;
use	base qw(CompanyIO);

__PACKAGE__->mk_accessors(qw(companies));

sub initialize {
	my	($self) = @_;
	$self->type('r');
	$self->SUPER::initialize();
}

sub read {
	my	($self) = @_;
	my	$csv = $self->csv();
	my	$file = $self->file();
	my	%cache = ();
	my	@companies = ();
	while (1) {
		my	$company;
		my	$data;
		my	($name,$value);
		my	$columns = $csv->getline($file);
		last if ($csv->eof());
		die $csv->error_diag()."(".$csv->error_input().")\r\n" unless ($csv->status());
		($name,$value) = split(/=/,shift(@$columns),2);
		if (defined($cache{$name})) {
			$company = $cache{$name};
		} else {
			$company = new Company();
			$company->name($name);
			$company->initialize();
			$cache{$name} = $company;
		}
		$data = $company->data();
		push(@$data,$value,@$columns);
		$#$data-- if ($data->[$#$data] eq '');
		$company->data($data);
	}
	@companies = values(%cache);
	$_->parse() foreach (@companies);
	$self->companies(\@companies);
	$self;
}


package	CompanyWriter;

use	strict;
use	Text::CSV;
use	base qw(CompanyIO);

__PACKAGE__->mk_accessors(qw(companies));

sub initialize {
	my	($self) = @_;
	$self->type('w');
	$self->SUPER::initialize();
}

sub write {
	my	($self) = @_;
	my	$file = $self->file();
	my	$csv = $self->csv();
	foreach my $company (sort { $a->name() cmp $b->name() } @{$self->companies()}) {
		$csv->combine($company->name(),@{$company->parsed()}) || die "invalid data.";
		$file->print($csv->string()."\r\n");
	}
}


package	main;

use	strict;

my	$reader = new CompanyReader();
my	$writer = new CompanyWriter();

if (scalar(@ARGV) != 2) {
	print "usage: ${0} IN_FILE OUT_FILE\r\n";
	exit;
}

$reader->file_name($ARGV[0]);
$reader->initialize();
$reader->read();

$writer->file_name($ARGV[1]);
$writer->initialize();

$writer->companies($reader->companies());
$writer->write();

$reader->finalize();
$writer->finalize();

1;

http://search.cpan.org/~makamaka/Text-CSV-1.00/lib/Text/CSV.pm