Re: Any Oracle OJB Users in Production? -- Thomas or someone,Comments on Datatype?

[Chris Lewington <cl...@ebi.ac.uk>]

Hi,

re. the OJB-specific tables and mismatch with oracle:

firstly if you don't use ODMG (or even if you do but not DList etc and don't do object
locking via a DB table) then you can ignore virtually all the OJB-specific tables. Also
if you use your own sequence manager and not those that come with OJB (eg high-low etc)
then you don't need any of the tables to run your applications, only the junit test
cases. If you are happy to take the risk just go for it - our stuff works on oracle and
postgres without any OJB-specific tables in our schema and we use ODMG (but with
in-memory locking and no TX-aware collection types). OK, so we lose the ability to run
the junit regression tests, but to get things up and running quickly for what you want
to do then you could park the table/type tweaking for now. So quick summary:

if you a) do your own sequence management, b) use in-memory obj locking, c) don't use
the ODMG collections, and d) don't want to run the junit stuff (ie have a little faith
:-)) then you don't need any of the OJB internal tables. You lose the regression
testing, but even if you go for OJB seq management you only need one table so just drop
the other ones that give you grief.

Just a thought,

cheers,

Chris

Jason McKerr wrote:

> Can someone who knows more about the core tables read and comment as
> well?