5:50 p.m.
Background: I’m a business user. Dev wrote the fact model and the initial
rule set. I can’t configure the fact model, the agenda groups, or the java
code that fires the rules. I can change the rules within the .drl file. My
question has to do with those rules.
The application is run as a web service. Each call has just one Autopolicy
in its own stateless session. An Autopolicy has one or more Vehicles, and a
Vehicle has one or more Package of insurance coverages. Autopolicy objects
come in with a random float (RandomGenId) between 0 and 1. Based on that
random float we assign the object to an experiment group (RuleFlowId).
Based on the experiment group we then set the value of the packages. (See
the code below.)
Below, rules 0A and 0B work when I use ruleFlowId != “test” and ruleFlowId
!=“control” respectively, but they don’t work with the much more
straightforward approach of ruleFlowId == “control” and ruleFlowId == “test”
(respectively). Does anybody have an idea why?
Thank you!
-------------------------------------------
rule "control group"
agenda-group "Rules-Recommendation"
salience 999998
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId < 0.5)
then
modify($autoPolicy) {setRuleFlowId("control")}; // Default ruleFlowId is
“1”
logger.debug("control group set");
end
rule "test group"
agenda-group "Rules-Recommendation"
salience 999997
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId >= 0.5)
then
modify($autoPolicy) {setRuleFlowId("test")}; // Default ruleFlowId is
“1”
logger.debug("test group set");
end
rule "Rule 0A: Set Packages for Control Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "test")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0A fired");
end
rule "Rule 0B: Set Packages for Test Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "control")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0B fired");
end
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
6:11 p.m.
It may be because in Java you have to compare strings using .equals("");
So in your case it would be:
ruleFlowId.equals("test") == true
It's because Strings are objects, and == tests to see if the String is
literally the same object as the other String, not just a String with the
same value. != works because they aren't the same object.
Hope that works,
Rich.
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
6:21 p.m.
Going by the example provided, the code is a constraint on the left hand side … so
it's valid to use == or !=. It's not Java code, it's MVEL and is automatically
converted to use the appropriate .equals("") method on the Java objects.
You would only need to use .equals("") inside an eval or in Java code on the
RHS.
So the answer is something else … but I haven't worked out what yet. :)
Steve
On 5 Feb 2013, at 17:11, rjr201 <rich.j.riley(a)gmail.com> wrote:
It may be because in Java you have to compare strings using
.equals("");
So in your case it would be:
ruleFlowId.equals("test") == true
It's because Strings are objects, and == tests to see if the String is
literally the same object as the other String, not just a String with the
same value. != works because they aren't the same object.
Hope that works,
Rich.
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
6:33 p.m.
Okay … well, without seeing what's been inserted as an Autopolicy in the first place,
I reckon the best plan is to debug it. It's probably worth changing to something like
this:
when
$autoPolicy : Autopolicy($ruleFlowId: ruleFlowId)
then
logger.debug("Rule flow ID = " + $ruleFlowId);
end
That should show you what value is held in the Autopolicy ruleFlowId property.
If you're comparing using != then the value could be just about anything. It could be
a String with a value other than "test" or "control". It could also be
null or a non-String object. When you use == "test", then you are saying that
the value must be a String with the value "test".
Steve
On 5 Feb 2013, at 17:21, Stephen Masters <stephen.masters(a)me.com> wrote:
Going by the example provided, the code is a constraint on the left
hand side … so it's valid to use == or !=. It's not Java code, it's MVEL and
is automatically converted to use the appropriate .equals("") method on the Java
objects.
You would only need to use .equals("") inside an eval or in Java code on the
RHS.
So the answer is something else … but I haven't worked out what yet. :)
Steve
On 5 Feb 2013, at 17:11, rjr201 <rich.j.riley(a)gmail.com> wrote:
> It may be because in Java you have to compare strings using .equals("");
>
> So in your case it would be:
>
> ruleFlowId.equals("test") == true
>
> It's because Strings are objects, and == tests to see if the String is
> literally the same object as the other String, not just a String with the
> same value. != works because they aren't the same object.
>
> Hope that works,
> Rich.
>
>
>
> --
> View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
> Sent from the Drools: User forum mailing list archive at Nabble.com.
> _______________________________________________
> rules-users mailing list
> rules-users(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
12:28 p.m.
Ah right, that's good to know.
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
6:41 p.m.
New subject: Can get "not equal" (!=) to work but not "equal" (==) in LHS
Are you setting the focus on the agenda group appropriately?
Remember that, with lock-on-active, you'll also have to TOGGLE the focus....
I tested the rules on 5.5.1-snapshot and everything seems to work,
which version are you using?
Davide
On 02/05/2013 09:50 AM, BenjaminWolfe wrote:
Background: I’m a business user. Dev wrote the fact model and the
initial
rule set. I can’t configure the fact model, the agenda groups, or the java
code that fires the rules. I can change the rules within the .drl file. My
question has to do with those rules.
The application is run as a web service. Each call has just one Autopolicy
in its own stateless session. An Autopolicy has one or more Vehicles, and a
Vehicle has one or more Package of insurance coverages. Autopolicy objects
come in with a random float (RandomGenId) between 0 and 1. Based on that
random float we assign the object to an experiment group (RuleFlowId).
Based on the experiment group we then set the value of the packages. (See
the code below.)
Below, rules 0A and 0B work when I use ruleFlowId != “test” and ruleFlowId
!=“control” respectively, but they don’t work with the much more
straightforward approach of ruleFlowId == “control” and ruleFlowId == “test”
(respectively). Does anybody have an idea why?
Thank you!
-------------------------------------------
rule "control group"
agenda-group "Rules-Recommendation"
salience 999998
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId< 0.5)
then
modify($autoPolicy) {setRuleFlowId("control")}; // Default ruleFlowId is
“1”
logger.debug("control group set");
end
rule "test group"
agenda-group "Rules-Recommendation"
salience 999997
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId>= 0.5)
then
modify($autoPolicy) {setRuleFlowId("test")}; // Default ruleFlowId is
“1”
logger.debug("test group set");
end
rule "Rule 0A: Set Packages for Control Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "test")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0A fired");
end
rule "Rule 0B: Set Packages for Test Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "control")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0B fired");
end
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
7:28 p.m.
The problem lies with your use of "lock-on-active" with every rule in
agenda-group "Rules-Recommendation". Once one of the rules "* group"
fires,
no more activations of other rules in that group are permitted; hence, the
effect of modify to set RuleFlowId is lost.
Activations that were around *before* a "* group" fires, are in
effect, and that's why you see "Rule 0*" firings when they use
"!=",
which happily return true due to whatever ruleFlowId was set to.
Discard all of that salience and lock-on-active tangle. Constrain
facts according to the initial values (the data modelers should have
provided reasonable settings, I'm assuming null here down below) of
the properties the facts have before you set them, e.g.,
$autoPolicy : Autopolicy(randomGenId < 0.5, ruleFlowId == null )
$autoPolicy : Autopolicy(randomGenId >= 0.5, ruleFlowId == null )
$autoPolicy : Autopolicy(ruleFlowId == "control")
$autoPolicy : Autopolicy(ruleFlowId == "test")
and watch the Engine doing good.
Cheers!
Wolfgang
On 05/02/2013, BenjaminWolfe <benjamin.e.wolfe(a)gmail.com> wrote:
Background: I’m a business user. Dev wrote the fact model and the
initial
rule set. I can’t configure the fact model, the agenda groups, or the java
code that fires the rules. I can change the rules within the .drl file. My
question has to do with those rules.
The application is run as a web service. Each call has just one Autopolicy
in its own stateless session. An Autopolicy has one or more Vehicles, and a
Vehicle has one or more Package of insurance coverages. Autopolicy objects
come in with a random float (RandomGenId) between 0 and 1. Based on that
random float we assign the object to an experiment group (RuleFlowId).
Based on the experiment group we then set the value of the packages. (See
the code below.)
Below, rules 0A and 0B work when I use ruleFlowId != “test” and ruleFlowId
!=“control” respectively, but they don’t work with the much more
straightforward approach of ruleFlowId == “control” and ruleFlowId == “test”
(respectively). Does anybody have an idea why?
Thank you!
-------------------------------------------
rule "control group"
agenda-group "Rules-Recommendation"
salience 999998
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId < 0.5)
then
modify($autoPolicy) {setRuleFlowId("control")}; // Default ruleFlowId is
“1”
logger.debug("control group set");
end
rule "test group"
agenda-group "Rules-Recommendation"
salience 999997
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(randomGenId >= 0.5)
then
modify($autoPolicy) {setRuleFlowId("test")}; // Default ruleFlowId is
“1”
logger.debug("test group set");
end
rule "Rule 0A: Set Packages for Control Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "test")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0A fired");
end
rule "Rule 0B: Set Packages for Test Group"
agenda-group "Rules-Recommendation"
no-loop true
lock-on-active true
dialect "mvel"
when
$autoPolicy : Autopolicy(ruleFlowId != "control")
$vehicles : Vehicles()
$packages : Packages()
$vehicle : Vehicle() from $vehicles.Vehicle
then
// set $packages
logger.debug("Rule 0B fired");
end
--
View this message in context:
http://drools.46999.n3.nabble.com/Can-get-not-equal-to-work-but-not-equal...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
4097
days inactive
4098
days old
6 comments
5 participants
participants (5)
-
BenjaminWolfe -
Davide Sottara -
rjr201 -
Stephen Masters -
Wolfgang Laun