Oana,
Although your example used a superclass/subclass
relationship, the problem is not due to that relationship, You could just
as well have written two CAR rules :
"if a car's price is <20000, then risk of
theft is low"
"if a car has the stealth radar avoidance option,
then risk of theft is high"
You might be upset by the infinite loop, but there is no
way to know how you would like to assess the theft risk of a car that satisfies
BOTH conditions.
Here are two ways you might deal with this.
a) maybe neither of these conditions is really
sufficient to determine the theft risk, but simply to influence that risk.
The "then" part of the rule should just increment/decrement a value the
accumulates the net risk. But don't have the increment/decrement also tell
drools that the CAR was modified.
b) maybe the rules are TOTALLY ORDERED in your analytical
model, and the first rule that applies is the only one that matters. You
can use SALIENCE in drools to capture ordering. If all the rules are in a single
activation group, then only the highest salience match will actually run (at
least, that is my understanding). Again, you must not tell drools that the
CAR has been modified just because you set its theft
assessment.
Hi everybody,
I have a class Car which is superclass for class
ConvertibleCar, two rules (described below) and a ConvertibleCar fact in working
memory that causes both rules to fire, resulting this way an unwanted change to
the car`s attribute and an infinite loop.
How do I evitate this problem:
the specialization class fact to not match its superclass column in a rule?
ConvertibleCar car =new ConvertibleCar(10000);
rule
"rule-1"
when
$car:Convertible()
then
$car.setPotentialTheftRating("high");
end
rule
"rule-2"
when
$car:Car(price<20000)
then
$car.setPotentialTheftRating("low");
end
Best regards,
Oana
The fish are biting.
Get
more visitors on your site using Yahoo!
Search Marketing.