> > db = SQLAlchemyFixture() > > # anytime before db.data().setup() ... > > db.session = my_session > > In my current system I have a single global session that is used for > everything. Is there any reason you can see that I could not just > reuse this session in all the test cases or should I be creating a new > on each time?
Are you using postgres? The only problem I foresee is if your test does some work with table instances shared by the fixtures but doesn't explicitly call rollback when there is an exception. You will probably even get a deadlock if that happens. `ps aux | grep postgres` will show if a fixture's delete statement is waiting on another transaction. I've tried to accomodate for this scenario so let me know if you run into it again so I can add some tests for it. > The other thing you lose with an SQL dump is that the output may not > work across different database backends. That is why I would really > like the loading of the table to be routed back through SA so we can > have some support for moving the testing data to whatever db's you end > up needed. (in my particular case this isn't really going to work > because I need GIS support which is non portable, but it sounds like a > nice capability to me) > agreed. I'm not entirely convinced that XML is the way to go, but this makes sense (and I like the plugin idea). It would be mighty fast with lxml.etree.iterparse(). still thinking... > ...and points out a problem I am going to > have using it. In my current code I am not keeping the tables or > mappers around. Instead I have a database manager class that sets > everthing up and simply holds onto the session and engine that should > be used to query the database. I rely upon the mapped classes to keep > track of the table and metadata references internally. > > So... is there any way to associated a dataset with the Class type > that is associated with the data in the dataset? > > for example maybe something like: > > class anything_I_want(DataSet): > mappedType = MyDataClass > class click: > name="click" > > or something else along these lines. This seems like it would work > well to tie the data back to the class type that is actually being > mapped. Then the "anything_I_want" class is really just a list of > MyDataClass objects that need to be populated into the database. Yes, I've made an attempt to support mapped classes but there are many different ways to map classes so I might not have captured them all. Please try: class anything_i_want(DataSet): class Meta: storable=MyDataClass class click: name="click" and let me know if that doesn't work, so I can get an idea for how you are using mapped classes. -Kumar --~--~---------~--~----~------------~-------~--~----~ 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 -~----------~----~----~----~------~----~------~--~---