we're going to have charcoal colored shirts.

i asked for sparkles for color, but might not happen.

schwern brought donuts!

i said i would improve on SQL by eliminating NULLS

someone pitched samy.pl/sql++/

our.coldhardcode.com/svn...

(that's the URL Jay is referencing)

we just started the podcast, so time to say naughty things!

object: is a group of columns, (a table), that has indexes

BEER is our example: beer has many distributors, with a brewer, and a name!

can add indexes add_unique_indexes()

belongs_to is the opposite end of has_many

has_one, might_have means that, and many_to_many is pretty complicated

create $schema from the Beer::Schema, and then ->connect($dsn) just like Class:BI

err. Class:\BI

"Chained Result Sets are what DBIC Great"

the code was bigger

create_ddl_dir() creates table statements

$schema->deploy() - creates database for you!

now for the tests!

the big thing to understand going from Class:BI to DBIx::Class

is that everything from DBIx::Class is a Result Set!

in the top level Schema - lots of ppl put in an init_schema() function to parse YAML, set things up, etc.

this results in much shorter tests! (testing with SQLite, instead of DB Brand X)

live debugging!

showing some code examples using beer from YAPC Asia

thx for the fix!

DBIx::Class will create the foreign key refs automatically

there were some weird things -- i think the more interesting case is what cheeseinspector did at PDXPUG

where you're taking an ORM and modeling a DB that already exists. i think that's much harder.

DBIx::Class::QueryLog::Analyzer - useful for debugging!

caching q: example of an attribute held in the object, versus pulling from the DB - you can force a recall at any point.

jay recommends doing everything in a transaction.

$schema->txn_do( blah blah );

(that's how to do a transaction)

DBIx::Class::Schema::Loader

one liner from the docs:

perl -MDBIx::Class::Schema::Loader=make_schema_at,dump_to_dir:./lib \

-e 'make_schema_at("My::Schema", { debug => 1 }, [ "dbig:dbname=foo","postgres" ])'

try Carp::REPL - read-eval-print-loop on die and or warn

time for practical beer!