Comments on: DBIx::Class vs Class::DBI vs Rose::DB::Object vs Fey::ORM

By: Grrrr

Grrrr — Mon, 29 Jun 2009 12:29:24 +0000

I would like to join the party and pitch an absolutely non-ORM solution, DBIx::Perlish, which comes up with a rather unusual way of doing the job. Rewriting the fetch example
from the SQL::DB pitch above:

my @items = db_fetch {
my $p : persons;
my $a : addresses;

join $p db_fetch {
$p->address == $a->id;
};
$a->city eq “Springfield”;
$p->age > 40;

sort desc => $p->age;
return $p->name, $p->age, $a->city;
limit: 10;
};

And rewriting the update:

db_update {
my $p : persons;

$p->name eq “Homer”;

$p->address = 2;
};

[disclaimer: I am the author of DBIx::Perlish]

By: Mark Lawrece

Mark Lawrece — Thu, 25 Jun 2009 05:57:52 +0000

Something that might also be of interest (but not necessarily for production use, yet) is SQL::DB:

http://search.cpan.org/~MLAWREN/SQL-DB/lib/SQL/DB.pm

This is something slightly less of an ORM than the above described packages, but lets you get very close to SQL for fetching:

1
2
3
4
5
6
7
8
9
10
11
12
13

my $iter = $db->fetch(
select => [$persons->name, $persons->age, $addresses->city],
from => $persons,
left_join => $addresses,
on => $addresses->id == $persons->address,
where => ($addresses->city == 'Springfield') & ($persons->age > 40),
order_by => $persons->age->desc,
limit => 10,
);

while (my $item = $iter->next) {
print $item->name, '(',$item->age,') lives in ', $item->city, "\n";
}

And for modifications…

1
2
3
4

$db->do(
update => [$persons->set_address(2)],
where => $persons->name == 'Homer',
);

[disclaimer: I am the author of SQL::DB]

By: Marlon

Marlon — Mon, 22 Jun 2009 20:13:55 +0000

You’re kidding right? Last time I checked you also “wrote” to databases.

By: gus

gus — Fri, 19 Jun 2009 23:17:13 +0000

a more in depth comparison is here: http://www.perlmonks.org/?node_id=700283

By: gus

gus — Fri, 19 Jun 2009 23:14:50 +0000

marion- isn’t that what memcache is for?

By: Marlon

Marlon — Fri, 19 Jun 2009 18:51:38 +0000

You really should run the benchmarks of Rose::DB vs DBIC. Your database layer is usually your slowest in your entire application, so that’s important. RDBO is also the ORM we use at CBSSPORTS.com because speed is important at this level. I’ll be interested in looking at what the new revision of DBIC with Moose does in that arena. On a side note, I agree that the generative queries are excellent on DBIC.

_Marlon_

By: John

John — Fri, 19 Jun 2009 17:19:07 +0000

The first CPAN release of Rose::DB::Object was on 2005-03-09, making it 4 years, 3 months, and 10 days old today. This version (0.01) has since been delete from CPAN, but you can see it on BACKPAN:

http://backpan.perl.org/authors/id/J/JS/JSIRACUSA/

and you can also see a record of it in the Changes file. (RDBO also existed in (then) CVS for about a year prior to its first CPAN release.)

By: Matt S Trout

Matt S Trout — Fri, 19 Jun 2009 12:50:04 +0000

DBIxDM is very interesting and Laurent Dami helps the DBIC team maintain SQL::Abstract – but he also doesn’t manage to market it to save his life so very few people have heard of it.

Maybe somebody should convince ldami to join ironman …

By: Zbigniew Lukasiak

Zbigniew Lukasiak — Fri, 19 Jun 2009 07:45:42 +0000

I understand that you had to make a selection here but I think you should include DBIx::DataModel. In my opinion it is a serious contender.