I was thrilled to see how nicely my model transforms into relational view, except one page has an objectified fact type and one subtype of a a shadowed OFT from another page that do not appear in the relational view.
OFT's on other pages diagram beautifully.
Any ideas what could cause this?
Kevin M. Owen
The relational view available in the current version is only an early proof-of-concept of what we are planning to support. It is not synchronized to the our DDL code generation; the tables, columns, and constraints it shows may or may not appear in the actual DDL.
We should have a release available by the end of this week that will contain a preview of our new relational mapping engine. Although for this preview release the relational schema will only be visible in the Model Browser (there won't be a diagram for it), what is shown will directly match what the DDL output will be.
In the mean time, the best suggestion I can offer is to try marking the objectified fact type as independent (through the Properties window).
Ok - that makes me feel better. I thought I broke something.
BTW, the entities in question all have IsIndependent option greyed out.
Hmm, given that it isn't IsIndependent that is the problem, I'm not sure off hand what would be causing those elements to not appear in the relational view.
Looking at the original post, is there a chance that the shadowed OFT was in the ER already? Shadowed, meaning established elsewhere in the model, the ER rep shouldn't be repeated for each invocation of the object in the ORM model diagram, only the original instance. I mentioned this because the OP'er is talking in terms of ORM model pages. Also, do OFT's always have their own representation in an ER diagram? I haven't thought this through, but I'd think that it might depend on how the OFT was referenced in the ORM diagram.
Kevin M. Owen
You are correct in that the number of diagrams / pages that any element of the model appears on does not effect how that element is mapped.
Objectified fact types may or may not have their own representation after they have been mapped. As with any other object type, it depends on the roles that they play. It is entirely possible for OFTs to be completely collapsed through the relationships that they participate in, and to therefore not cause a key-total table to be generated for them.
Although it is tricky for a beginner to get an OFT (or other fact types, for that matter) to be mapped to its own table instead of being absorbed (collapsed), I have had some success in doing this, although it's not yet clear to me in my mind what factors really determine this outcome. I reckon I will get better insight with more experience.
Most of my OFT's do appear in the relational view, although the one I was referring to still stubbornly refuses to do so. I don't see the subtle difference from my other OFT's that suppress it in the RV. However, after confirming that it was not absorbed into another entity, I confirm that it does indeed output to its own table in the DDL, just as Kevin said.