[rules-users] not exists / forall(not) logically equivalent, behave differently
Edson Tirelli
tirelli at post.com
Mon Oct 6 14:25:30 EDT 2008
I believe there is an open ticket for a bug when using "not(exists())"
in 5.0. I did not investigated yet, but I believe this is a regression in
the new asymmetric matching algorithm in 5.0. Anyway, we will fix it before
releasing the final version.
Meantime, just use:
not( Tail(dog == $dog) )
Since the "exists" is redundant there because "not" is the existential
operator, not the "boolean" operator.
[]s
Edson
2008/10/6 tim tim <timbaermannextra at googlemail.com>
> hello there,
>
> i have a rule, that should fire, when a required part of a whole is
> missing.
> lets say the whole is a dog, and the required part the tail.
> then the rule should fire, when there is a dog without a tail.
>
> i stated the rule in two logically equivalent ways.
>
> 1) when we have a dog, and all existing tails belong to other peoples dogs,
> the rule should fire:
>
> rule "no dog without a tail; forall"
> when
> $dog :Dog()
> forall($t : Tail()
> Tail(dog != $dog))
> then
> // fire rule
> end
>
>
> 2) when we have a dog and no tail exists that belongs to this dog the rule
> should fire:
>
> rule "no dog without a tail; not exists"
> when
> $dog :Dog()
> not( exists( Tail(dog == $dog)))
> then
> // fire rule
> end
>
>
> and they really do behave similar, but not the same..
> workingMemory.insert(dog);
> workingMemory.insert(dog.getTail());
> workingMemory.fireAllRules();
>
> // here both did not fire, but when i go on:
>
> workingMemory.modifyRetract(dogHandle)
> workingMemory.modifyInsert(dogHandle, dog)
> workingMemory.fireAllRules();
> // now the not(exists(..)) version did fire, whereas the forall(..) version
> did not.
>
>
> of course both rules should not have fired, nothing changed for the dogs
> relationship
> with its tail.
>
> i am using the 5.0.0.snapshot version of drools but i can not imagine that
> this is an
> unknown bug, as this is rather basic predicate logic.
> what am i missing?
>
> best, tim
>
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
--
Edson Tirelli
JBoss Drools Core Development
JBoss, a division of Red Hat @ www.jboss.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20081006/f4843a80/attachment.html
More information about the rules-users
mailing list