Arnon, I'm sure you are well-meaning, and I appreciate your enthusiasm, but at this point, I think we're going to have to move on. I was encouraged when you started providing examples and we had something concrete to discuss, and I think that has clarified some issues, but it appears this is no longer productive.
You began by proposing that we "merely" build something like the SQLA ORM. Now that even SQLA doesn't meet your needs, you have progressed to requesting a hypothetical ORM that would not only be substantially more sophisticated than even SQLA but also, it appears, logically impossible to implement (at least in the general case). You seem to think you can query the session without having a way to uniquely associate objects with their RDBMS records, and that you can create dummy partial-objects and somehow associate and merge them with real database records after the fact. And all this for what will likely be modest performance improvements in limited use cases where alternative strategies would be much simpler to implement. Since you do know how to program and have apparently worked out the logical details of your desired functionality, I would suggest you should therefore be able to implement at least a basic proof-of-concept on your own. Maybe see how that goes and get back to us. Anthony -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.