just chiming in as a company that is itching pretty hard for this right now, our use cases and concerns
> yeah...i think most people want to just use bind_to with their > session in order to switch around databases for now....its the most > expedient. this would solve most of our problems except that we don't the ORM bits at all and as such don't access sessions through a session object (i may be wrong about how this works, too, so feel free to tell me), the options I see along this route for us is using a DynamicMetadata and calling setEngine whenever we want to point at other places, but living in an async world switching global stuff around feels inherently bad. a use case like this would work fine for us everywhere i can thnk of: # selecting user "jimmy" # get_engine would return the engine we want for a particular user u = user.select().execute(nick="jimmy", engine=get_engine("jimmy")) > the basic idea is that the Connection object would be replaced by a > proxy which intercepts SQL and applies rules to it, which results in > any number of activities...the two that come to mind is "execute this > SQL on database X", the other is "execute this SQL on databases X, Y, > and Z", where the latter is used to build instant clustered writes. i'm in the same mind that i wouldn't actually trust non-database to handle the clustered writes (though i see the uses for pet projects), but what would be useful is to have a pool of engines for reads, doesn't even need to do any load balancing just have some backups in case a connection fails for any reason. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~----------~----~----~----~------~----~------~--~---