I haven't come across it as yet. I wouldn't expect that the performance issues would be related specifically to using it in drools. Even if the rule engine is getting properties over and over for conditions, only once needs to trigger the lazy load, so you would notice a hit the first time, and it just happens to be in assertObject etc that makes it look like its the rule engine slowing down.
If its hibernate 3 and up, its pretty agressive in being lazy (how can you be "agressive" and "lazy" at the same time ;) so often you need to turn it off in the mappings, or change the query to have fetch instructions - might be worth a try when tuning. Michael. On 2/2/06, Geoffrey Wiseman <[EMAIL PROTECTED]> wrote: > > This seems to have impacted our performance at least somewhat, so if there > is good knowledge on the best way to optimize use of Drools on lazy-loaded > domain objects via Hibernate, please, fire them off to the list so that I > can pass them on. > > On 2/1/06, Felipe Piccolini <[EMAIL PROTECTED]> wrote: > > > > Michael, > > > > Yes, lazy loading, and yes... proxies as usual with hibernate. > > > > But I think I solved the problem, because there was some rules with > > tuples (the same subject you are discussing all this time), two facts > > of the same type asserted with no condition asking the different > > instance. > > > > Thanks anyway :) > > > > Tuesday, January 31, 2006, 6:59:45 PM, you wrote: > > > > > lazy loading going on? Are they hibernate proxies? > > > > > On 1/31/06, Felipe Piccolini <[EMAIL PROTECTED] > wrote: > > > Hi all, > > > > > I have a problem asserting some complex objects. Im asserting > > > persistent objects with hibernate. This objects are very complex > > > and in some cases assering one of them takes 3 secs!!... Im > > > working with 500 rules aprox and this objects are parameters for > > > just 50 or 100 of them. > > > > > What could be the problem?... > > > > > ---------------------- > > > Felipe Piccolini > > > [EMAIL PROTECTED] > > > > > > > > > > > > > > > > -------------------------- > > Felipe Piccolini > > [EMAIL PROTECTED] > > > > > > > -- > Geoffrey Wiseman > >
