> On Fri, 18 Apr 2008 18:04:38 +0200, Frans Bouma <[EMAIL PROTECTED]> wrote:
> >> hmmm....seems a little wierd, it would be nice to query and bind to Dogs
> >> directly (if you see what I mean).
> >>
> >> So I create a form and grid and bind to the Animals base Table (as it's
> >> the only one I've got and remove all the stuff I'm not interested
> >> in.....add something to the grid and of course it defaults to the base
> >> type.
> >
> >        you can forget databinding if inheritance is involved: databinding
in
> >grids works with a single set of properties, and typically grids pick
either
> >the first entry in the bound set to determine these or ask the ITypedList
> >implementation of the set, if available.
> >
> >        if you have 2 subtypes of animal: Dog and JellyFish, both will have
> >their own unique properties: what should a grid do: display these columns
or
> >not? If so, what should be happening when a row represents a dog and a
column
> >specific for jellyfish is changed?
>
> This is why I thought it wierd that the inheritance mapping doesn't expose
> the subtypes as 'virtual' tables, then you would bind to them directly,
> the mapping would work to map this to the database and everyones
> happy....the client gets an OO model the database an RDb one....to me this
> is the point of the O/R mapping....as I say I think this implementation is
> half baked.

        I think this is a limitation inside their o/r system. I'm currently
writing a set of templates for our code generator engine to use our meta-data
to spit out linq to sql files (as our designer has model management for
database changes, their designer doesn't, so why not ;)) and I indeed saw that
there aren't any Table<subtype> typed properties for my subtypes on the
datacontext. I haven't tested yet what happens if I do add them manually to
the datacontext class, if things go bezerk during fetching or not.

        They also store all discriminator values/types combinations at the
hierarchy root, so I think bending their framework to use the subtypes as
normal types would probably not work...

                FB

>
> >                FB
> >
> >===================================
> >This list is hosted by DevelopMentorR  http://www.develop.com
> >
> >View archives and manage your subscription(s) at
> http://discuss.develop.com
>
> ===================================
> This list is hosted by DevelopMentorR  http://www.develop.com
>
> View archives and manage your subscription(s) at http://discuss.develop.com

===================================
This list is hosted by DevelopMentorĀ®  http://www.develop.com

View archives and manage your subscription(s) at http://discuss.develop.com

Reply via email to