On Aug 17, 2011, at 1:06 PM, Tim Black wrote: > On 08/17/2011 11:16 AM, Michael Bayer wrote: >> >> Great, and to answer your TODO no, metadata.bind = engine is not necessary >> and I try to discourage people from using that activity by default (but the >> plethora of Turbogears and Pylons tutorials that keep using it makes this >> task more or less impossible). A discussion of that feature specifically >> is at >> http://www.sqlalchemy.org/docs/core/schema.html#binding-metadata-to-an-engine-or-connection >> . > I read that, and because the docs indicate removing metadata.bind = engine1 > would make me lose "implicit" execution (unless it's provided by some other > means in TurboGears' config), I worried that might break TurboGears' > transaction management or query execution configuration somehow--is that > possible? I only ask because TurboGears' creators may have some question > like this in mind--maybe you can see better than I whether this indicates a > way the docs can be improved. After reading more of the SQLAlchemy docs, it > appears to me the answer to that question is "No," because "implicit" > execution and transaction management are two very different and unrelated > things in SQLAlchemy, and that whatever "implicit" execution TurboGears may > need is probably provided by binding the engine to the session. And, so far, > my code works without the metadata.bind = engine1 statement. > > I'll mention one more thing where maybe the docs should be improved--when I > created the lines below, I had to read through SQLAlchemy's code to try to > figure out whether I could use both bind (singular) and binds (plural) > simultaneously on the same object. Initially the docs made me feel > uncomfortable using both bind and binds; it seemed like if bind stores one > connectable in the session, binds might replace that one connectable with a > list of connectables, so maybe I could only use one and not both (bind and > binds). But, through trial and error (which docs need never attempt to > completely replace), using both works! > > DBSession.configure(bind=engine1) > - DBSession2.configure(bind=engine2) > + DBSession.configure(binds={Ticket:engine2}) > > But here is how the docs could be made more clear - simply state 1. whether > it is allowable to use bind and binds simultaneously on the same object, and > (if 2. is the case) 2. that the engine specified in bind will be used for all > classes, mappers, and tables except where another engine is specified by > binds. For brevity's sake, I note that 1. is implicit in 2.
well the docs so far are in http://www.sqlalchemy.org/docs/orm/session.html#sqlalchemy.orm.session.Session.get_bind and basically it will try to return the most specific "bind" first. ie.....I was about to start typing the order here. What's the use in that, when I can type it into the docs instead, so click over there to see the exact order of resolution. -- 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 sqlalchemy+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.
