[Dbix-class] RFC: DBIx::Class::TypedView (or Is this CPAN worthy?)

Mark Hedges hedges at ucsd.edu
Wed Jan 10 20:51:01 GMT 2007



On Sat, 2 Dec 2006, Matt S Trout wrote:
> 
> On 2 Dec 2006, at 15:28, Nilson Santos Figueiredo Junior wrote:
> 
> > On 12/2/06, Matt S Trout <dbix-class at trout.me.uk> wrote:
> > > First thing: inflating to different classes based on a column
> > 
> > Actually, when selecting it doesn't really inflate to the different
> > classes. In the table A, table B, view V example above, when selecting
> > you'd get V objects, not A or B objects.
> > 
> 
> Right, then what we definitely want is a smarter ResultSource object - one
> that produces the appropriate union select when ->from is called, and can
> redirect insert/update requests appropriately. the problem, as I say, is that
> currently the ResultSet passes the source straight through to storage whereas
> what we really need is to be passing the request to the source first so it can
> indirect through sources for tables A and B as appropriate.
> 
> Might be worth having a look at John's bytea-related branch and seeing how
> much trouble it'd be to do this - then it could be completely transparent to
> the class itself ...

I wondered if anyone was still interested in this concept and 
doing work on it.  I might have an application for it at a new job. 

Am I wrong that this idea is there are multiple identical tables 
for the same class of object, and the search methods build 
unions across the tables and send updates to the correct tables?

Mark



More information about the Dbix-class mailing list