Re: [Modeling-users] instantiating objects from (aka, mapping entities to) stored procedure result sets

[Sebastien B. <sbi...@us...>]

Ezra E. <sf-...@pr...> wrote:
> Hi,
>=20
>    Looking through the docs it seems that the modeling framework makes so=
me
> stronger assumptions about how entities are mapped to database result sets
> than one finds in EOF.  In particular, is there a way to map objects to t=
he
> results of stored procedures?  (Again, how hard would it be for an EOF-sa=
vvy
> developer to add?)

Well, you're right, we're making stronger assumptions than in EOF.

Unfortunately, here again, I've no experience at all w/ stored proc. so
it's hard to tell how much time it would take, just because I cannot
foresse any possible difficulty that may arise.

  However, it's also no clear to me what you really mean by <<mapping
  objects to the results of stored proc.>>. Does this mean that such
  entities will be read-only, or should there also be some procedures
  to insert & delete?

BTW, if I assume that stored procedure is not a too ugly beast and that,
from a python db-adaptor point of view it does not differ a lot from the
standard SQL manipulating selects/inserts/deletes (sounds like a
middle-to-strong assumption, but never mind), then I guess this won't be
to hard to implement. MDL.Entity should probably store the stored
procedures to be used. Having a quick look at the code, I think that
this info. about stored proc. to use can be used, at best, during the
DBContext's performChanges() phase, when AdaptorOperation are built from
DatabaseOperation.

  Does this answers the question? Probably not fully ;)  Again, please
  fill a RFE at sourceforge if it's something you're likely to need
  quickly. But the story is that I have the feeling that it should not
  be a difficult thing to add to the framework.

Last, I'd say as usual that, if you have small toy-examples that we can
play with, that would be great --at least for me who never played w/ any
stored proc.

-- S=E9bastien.