[Dbix-class] first Muldis DB example code (was Re: SQL Generation with SQL::DB)

[Darren Duncan]

At 2:48 AM -0700 9/17/07, Darren Duncan wrote:
>     my $r1_ast = [ 'Relation', 'sys.Core.Relation.Relation', [
>         {
>             'x' => [ 'PInt', 'perl_pint', 4 ],
>             'y' => [ 'PInt', 'perl_pint', 7 ],
>         },
<snip>
>     my $r2_ast = [ 'Relation', 'sys.Core.Relation.Relation', [
>         {
>             'y' => [ 'PInt', 'perl_pint', 5 ],
>             'z' => [ 'PInt', 'perl_pint', 6 ],
>         },
<snip>
>     my $r3 = $dbms->call_func( 'sys.Core.Relation.join', {
>         'topic' => [ 'QuasiSet', 'sys.Core.Spec.QuasiSetOfRelation', [
>             $r1, # or $r1_ast
>             $r2, # or $r2_ast
>         ],
>     } );
>     my $r3_ast = $r3->fetch();
>   
>     # Then $r3_ast contains:
>     # [ 'Relation', 'sys.Core.Relation.Relation', [
>     #     {
>     #         'x' => [ 'PInt', 'perl_pint', 3 ],
>     #         'y' => [ 'PInt', 'perl_pint', 2 ],
>     #         'z' => [ 'PInt', 'perl_pint', 1 ],
>     #     },
<snip>
>That example declares two Perl variables containing rowset/table 
>data, then does a (N-ary) relational join (natural inner join) on 
>them, producing a third Perl variable holding the rowset/table data 
>of the result.

As an addendum, if you couldn't figure it out for yourself ...

If this example were against actual SQL tables mirroring $r1 and $r2, 
then the generated SQL for the above join would look like:

     SELECT DISTINCT r1.x AS x, r1.y AS y, r2.z AS z
     FROM r1 INNER JOIN r2 USING (y)

-- Darren Duncan