Aaron Mulder wrote: > We've used CMP a fair amount before at work, and IMO there's no reason > it can't be made to perform well for most cases (you know, not > necessarily for dynamic queries or updating 1000s of rows at once). > But I have to say, the SQL generation in Geronimo is something I've > never seen before, and it's a litte more cumbersome than I'd expect. > I think we should offer 2 options: one option to update everything > except the primary keys on every update (the "always use same > statment" option where it just says update x set foo=?, bar=?, baz=?, > etc.), and another option to only update the fields that were changed > on every update (may cause different SQLs for each update, but avoids > problems with triggers and so on). There can be a flag in the > deployment plan to say which strategy to use for each EJB.
Why would anyone select the first option, less than optimal single update? Dan.
