Felipe,
Note that in this situation it seems a waste as your evaluating conditions
for rules you have no intention of firing. Its better to do away with the
agenda filters and have seperate rulebases and seperate working memories and
just assert the same facts into both working memories.
Lionel
On 3/8/06, Lionel Port <[EMAIL PROTECTED]> wrote:
>
> Michael was not saying to modify the facts in the consequence. Instead, he
> was saying that you it would be _possible_ iterate through the facts after
> the rules had fired and modify them.
>
> You could do something like:
>
>
> memory.fireAllRules(agendaFilter1);
>
> // do something
>
> // modify all facts so that rules are re-added to agenda.
> List objects = memory.getObjects();
> for (Object fact : objects )
> {
> FactHandle factHandle = memory.getFactHandle(fact);
> memory.modifyObject(factHandle, fact);
> }
>
> memory.fireAllRules(agendaFilter2);
>
>
>
>
>
>
>
>
>
>
>
> On 3/8/06, Felipe Piccolini < [EMAIL PROTECTED] > wrote:
> >
> > Michael,
> >
> > But, how can you modify 'all' the objects on the working memory?..
> > when a consequence finds a drools.modifyObject(Fact) inside the .drl
> > all the rulebase is checked again no matter the next line...
> >
> > I want to do:
> >
> > <java:consequence>
> > drools.modifyObject (myFact1);
> > drools.modifyObject(myFact2);
> > </java:consequence>
> >
> > But the matter is the second line is never executed.
> >
> > Monday, March 6, 2006, 10:51:29 PM, you wrote:
> >
> > > no when its finished, the agenda is clear.
> >
> > > I think what you really want is "Agenda Groups" where it is partioned
> > based
> > > on some logic - but this is an upcoming drools 3 feature.
> >
> > > in the meantime, you can also iterate through and "modify" all the
> > objects
> > > that are already in working memory, causing rules to be put on the
> > agenda
> > > again (possibly).
> >
> > > However, I think the correct way is what you suggested, as you are
> > talking
> > > about 2 different points in your applications workflow, hence 2
> > rulesets.
> > > The facts just happen to be the same. In any case, the "cost" is not
> > so much
> > > in iterating and calling assertObject, but the work done to calculate
> > what
> > > rules should fire, which in any case needs to happen twice.
> >
> > > Someone may have another more efficient suggestion however.
> >
> > > On 3/7/06, Lionel Port < [EMAIL PROTECTED]> wrote:
> > >>
> > >> When you use the fireAllRules(AgendaFilter) method on the
> > WorkingMemory
> > >> does
> > >> it clear the Agenda when its done or can you fire different rules in
> > the
> > >> working memory at different points in time without modifying the
> > facts in
> > >> between.
> > >>
> > >> The rules I have can be broken into two distinct rulesets that need
> > to be
> > >> executed at different points but the conditions largely operate on
> > the
> > >> same
> > >> facts, so I was trying to save time by asserting all the facts once
> > and
> > >> then
> > >> doing something like:
> > >>
> > >> memory.fireAllRules(agendaFilter1);
> > >> // do something not related to drools
> > >> memory.fireAllRules(agendaFilter2);
> > >>
> > >>
> > >> It doesn't seem to be working though. Should I just put the rules in
> > >> different rule bases and assert the facts before I use it.
> > >>
> > >>
> >
> >
> > --------------------------
> > Felipe Piccolini
> > [EMAIL PROTECTED]
> >
> >
>
