10:28 a.m.
Hi all,
I try to detect the beginning of a scenario which is characterized by an
event happening after no other event of the same type has happened in the
last 90sec.
I can't understand why only the first of those two rules is working.
rule "Detect Scenario beginning"
when
$start : RawEvent()
not RawEvent(id == $start.id, $start after[0ms, 90s] this)
then
System.out.println($start);
end
rule "Detect Scenario beginning alt"
when
$start : RawEvent()
not RawEvent(id == $start.id, this before[0ms, 90s] $start)
then
System.out.println($start);
end
Moreover if the first sent event is a rawEvent none of them will detect it.
Where am i wrong ? (still a beginner with drools)
PS : I already posted it on the forum, but i forgot to complete all the
subscription steps so it was rejected. Sorry for the double post.
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
10:48 a.m.
You say that the first rule is working, but then you also write that
if the first event is a RawEvent neither one works. Isn't this a
contradiction?
According to the definition and using "t" for "timestamp", we have
A after[x,y] B
equalling, by definition
x <= A.t - B.t <= y
or
B.t + x <= A.t <= B.t + y
and with [x,y] set to [0,90s] we have
B.t <= A.t <= B.t + 90
(A before[x,y] B is the same except for exchanging the roles of A and B.)
However, any RawEvent R inhibits both rules from firing, since it matches
both the pattern after $start and the pattern after "not": its "id"
is
necessarily the same and clearly
R.t <= R.t <= R.t + 90
Try adding
not RawEvent( this != $start, id == $start.id,...
-W
On 1 September 2012 17:28, Rob21 <rob.perrez(a)gmail.com> wrote:
Hi all,
I try to detect the beginning of a scenario which is characterized by an
event happening after no other event of the same type has happened in the
last 90sec.
I can't understand why only the first of those two rules is working.
rule "Detect Scenario beginning"
when
$start : RawEvent()
not RawEvent(id == $start.id, $start after[0ms, 90s] this)
then
System.out.println($start);
end
rule "Detect Scenario beginning alt"
when
$start : RawEvent()
not RawEvent(id == $start.id, this before[0ms, 90s] $start)
then
System.out.println($start);
end
Moreover if the first sent event is a rawEvent none of them will detect it.
Where am i wrong ? (still a beginner with drools)
PS : I already posted it on the forum, but i forgot to complete all the
subscription steps so it was rejected. Sorry for the double post.
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
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
11:41 a.m.
Actually i meant that the first rule worked for every beginning of scenario,
except the first scenario if the first event is a RawEvent. The second rule
doesn't work at all and gives me this :
Thread [main] (Suspended (exception NullPointerException))
BeforeEvaluatorDefinition$BeforeEvaluator.evaluateCachedLeft(InternalWorkingMemory,
VariableRestriction$VariableContextEntry, Object) line: 343
EvaluatorConstraint.isAllowedCachedLeft(ContextEntry, InternalFactHandle)
line: 67
DoubleBetaConstraints.isAllowedCachedLeft(ContextEntry[],
InternalFactHandle) line: 168
NotNode.assertLeftTuple(LeftTuple, PropagationContext,
InternalWorkingMemory) line: 80
SingleLeftTupleSinkAdapter.doPropagateAssertLeftTuple(PropagationContext,
InternalWorkingMemory, LeftTuple) line: 196
SingleLeftTupleSinkAdapter.createAndPropagateAssertLeftTuple(InternalFactHandle,
PropagationContext, InternalWorkingMemory, boolean, LeftInputAdapterNode)
line: 145
LeftInputAdapterNode.assertObject(InternalFactHandle, PropagationContext,
InternalWorkingMemory) line: 154
CompositeObjectSinkAdapter.doPropagateAssertObject(InternalFactHandle,
PropagationContext, InternalWorkingMemory, ObjectSink) line: 497
CompositeObjectSinkAdapter.propagateAssertObject(InternalFactHandle,
PropagationContext, InternalWorkingMemory) line: 382
ObjectTypeNode.assertObject(InternalFactHandle, PropagationContext,
InternalWorkingMemory) line: 235
EntryPointNode.assertObject(InternalFactHandle, PropagationContext,
ObjectTypeConf, InternalWorkingMemory) line: 240
NamedEntryPoint.insert(InternalFactHandle, Object, Rule, Activation,
ObjectTypeConf) line: 337
NamedEntryPoint.insert(Object, boolean, boolean, Rule, Activation) line:
298
ReteooStatefulSession(AbstractWorkingMemory).insert(Object, boolean,
boolean, Rule, Activation) line: 888
ReteooStatefulSession(AbstractWorkingMemory).insert(Object) line: 847
StatefulKnowledgeSessionImpl.insert(Object) line: 269
DroolsTest.sendEvents(String, StatefulKnowledgeSession) line: 85
DroolsTest.main(String[]) line: 62
About "this != $start", i tried that also as the documentation contains a
similar example with heartbeats.
Still for this rule :
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, $start after[0s, 90s] this)
then
System.out.println($start);
end
The application terminates and Drools answers this :
Unable to Analyse Expression this != $start && after0.evaluate( $start, this
):
[Error: unable to resolve method using strict-mode:
com.sample.RawEvent.after0()]
[Near : {... this != $start && after0.evaluate( $start, this ....}]
^ : [Rule name='Detect it']
java.lang.IllegalArgumentException: Could not parse knowledge.
at com.sample.DroolsTest.readKnowledgeBase(DroolsTest.java:109)
at com.sample.DroolsTest.main(DroolsTest.java:54)
Which I don't understand
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
11:51 a.m.
And about the before design :
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, this before[0ms, 90s]
$start)
then
System.out.println($start);
end
I obtain this after a few events have been detected :
Thread [Thread-0] (Suspended (exception NullPointerException))
BeforeEvaluatorDefinition$BeforeEvaluator.evaluateCachedLeft(InternalWorkingMemory,
VariableRestriction$VariableContextEntry, Object) line: 343
EvaluatorConstraint.isAllowedCachedLeft(ContextEntry, InternalFactHandle)
line: 67
TripleBetaConstraints.isAllowedCachedLeft(ContextEntry[],
InternalFactHandle) line: 199
NotNode.retractRightTuple(RightTuple, PropagationContext,
InternalWorkingMemory) line: 183
ObjectTypeNode.retractObject(InternalFactHandle, PropagationContext,
InternalWorkingMemory) line: 283
ReteooWorkingMemory$WorkingMemoryReteExpireAction.execute(InternalWorkingMemory)
line: 514
ReteooStatefulSession(AbstractWorkingMemory).executeQueuedActions() line:
977
DefaultAgenda.fireUntilHalt(AgendaFilter) line: 1422
ReteooStatefulSession(AbstractWorkingMemory).fireUntilHalt(AgendaFilter)
line: 755
ReteooStatefulSession(AbstractWorkingMemory).fireUntilHalt() line: 731
StatefulKnowledgeSessionImpl.fireUntilHalt() line: 247
DroolsTest$rulesThread.run() line: 45
As you said, the before and after design should react the same way. I work
in stream mode with a pseudo clock i advance every time i insert an event to
the precise time the event to be inserted is dated. (If that can help in
anything)
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
12:13 p.m.
Both look like a bug to me. Even if RawEvent is an (abstract) base class,
the "unable to resolve method" should not occur. (Are you using dialect
MVEL?)
Are you using a separate thread running kSession.fireUntilHalt()? There was
an
alert indicating a race condition in 5.4.0.
Can you post a reduced set of Java and DRL code reproducing both errors?
-W
On 1 September 2012 18:41, Rob21 <rob.perrez(a)gmail.com> wrote:
Actually i meant that the first rule worked for every beginning of
scenario,
except the first scenario if the first event is a RawEvent. The second rule
doesn't work at all and gives me this :
Thread [main] (Suspended (exception NullPointerException))
BeforeEvaluatorDefinition$BeforeEvaluator.evaluateCachedLeft(InternalWorkingMemory,
VariableRestriction$VariableContextEntry, Object) line: 343
EvaluatorConstraint.isAllowedCachedLeft(ContextEntry,
InternalFactHandle)
line: 67
DoubleBetaConstraints.isAllowedCachedLeft(ContextEntry[],
InternalFactHandle) line: 168
NotNode.assertLeftTuple(LeftTuple, PropagationContext,
InternalWorkingMemory) line: 80
SingleLeftTupleSinkAdapter.doPropagateAssertLeftTuple(PropagationContext,
InternalWorkingMemory, LeftTuple) line: 196
SingleLeftTupleSinkAdapter.createAndPropagateAssertLeftTuple(InternalFactHandle,
PropagationContext, InternalWorkingMemory, boolean, LeftInputAdapterNode)
line: 145
LeftInputAdapterNode.assertObject(InternalFactHandle,
PropagationContext,
InternalWorkingMemory) line: 154
CompositeObjectSinkAdapter.doPropagateAssertObject(InternalFactHandle,
PropagationContext, InternalWorkingMemory, ObjectSink) line: 497
CompositeObjectSinkAdapter.propagateAssertObject(InternalFactHandle,
PropagationContext, InternalWorkingMemory) line: 382
ObjectTypeNode.assertObject(InternalFactHandle, PropagationContext,
InternalWorkingMemory) line: 235
EntryPointNode.assertObject(InternalFactHandle, PropagationContext,
ObjectTypeConf, InternalWorkingMemory) line: 240
NamedEntryPoint.insert(InternalFactHandle, Object, Rule,
Activation,
ObjectTypeConf) line: 337
NamedEntryPoint.insert(Object, boolean, boolean, Rule, Activation)
line:
298
ReteooStatefulSession(AbstractWorkingMemory).insert(Object,
boolean,
boolean, Rule, Activation) line: 888
ReteooStatefulSession(AbstractWorkingMemory).insert(Object) line:
847
StatefulKnowledgeSessionImpl.insert(Object) line: 269
DroolsTest.sendEvents(String, StatefulKnowledgeSession) line: 85
DroolsTest.main(String[]) line: 62
About "this != $start", i tried that also as the documentation contains a
similar example with heartbeats.
Still for this rule :
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, $start
after[0s, 90s] this)
then
System.out.println($start);
end
The application terminates and Drools answers this :
Unable to Analyse Expression this != $start && after0.evaluate( $start,
this
):
[Error: unable to resolve method using strict-mode:
com.sample.RawEvent.after0()]
[Near : {... this != $start && after0.evaluate( $start, this ....}]
^ : [Rule name='Detect it']
java.lang.IllegalArgumentException: Could not parse knowledge.
at com.sample.DroolsTest.readKnowledgeBase(DroolsTest.java:109)
at com.sample.DroolsTest.main(DroolsTest.java:54)
Which I don't understand
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
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
1:18 p.m.
I'm not using MVEL, but i do use fireUntilHalt in another thread.
Main Droolstest.java : http://pastebin.com/tmEcQKhi
RawEvent.java : http://pastebin.com/VSDhW5J5
BasicEvent.java : http://pastebin.com/0dx4QQyu
Sample.drl : http://pastebin.com/sW89mcVX
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
1:23 p.m.
I tried using FireAllRules each time i inserted an object instead of
FireUntilHalt in another thread. Unfortunately it results in the same error
message :
Unable to Analyse Expression this != $start && after0.evaluate( $start, this
):
[Error: unable to resolve method using strict-mode:
com.sample.RawEvent.after0()]
[Near : {... this != $start && after0.evaluate( $start, this ....}]
^ : [Rule name='Detect it']
java.lang.IllegalArgumentException: Could not parse knowledge.
at com.sample.DroolsTest.readKnowledgeBase(DroolsTest.java:106)
at com.sample.DroolsTest.main(DroolsTest.java:54)
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
2:39 a.m.
@Mario: Could this be something in mvel2-2.1.0.drools16.jar?
With Drools 5.3.0, this works fine:
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, $start after[0ms, 90s]
this)
then
System.out.println("detected: " + $start);
end
Moving on to 5.4.0, you get the error "Unable to Analyse Expression this !=
$start &&...", as reported previously by Rob.
@Rob: As a workaround, use
not $re: RawEvent(eval($re != $start),...
(And I'd recommend using new Random( 12345 ) to obtain repeatable test
outputs.)
-W
On 1 September 2012 20:23, Rob21 <rob.perrez(a)gmail.com> wrote:
I tried using FireAllRules each time i inserted an object instead of
FireUntilHalt in another thread. Unfortunately it results in the same error
message :
Unable to Analyse Expression this != $start && after0.evaluate( $start,
this
):
[Error: unable to resolve method using strict-mode:
com.sample.RawEvent.after0()]
[Near : {... this != $start && after0.evaluate( $start, this ....}]
^ : [Rule name='Detect it']
java.lang.IllegalArgumentException: Could not parse knowledge.
at com.sample.DroolsTest.readKnowledgeBase(DroolsTest.java:106)
at com.sample.DroolsTest.main(DroolsTest.java:54)
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
5:08 a.m.
Thanks it works fine with this !
Not sure if it will help you for the debug, but I also noticed using "$re"
in place of "this" with the "after" keyword gives the following (with
or
without the fireUntilHalt):
java.lang.ArrayIndexOutOfBoundsException: 1
at
org.drools.base.mvel.MVELCompilationUnit.getFactHandle(MVELCompilationUnit.java:423)
...
I don't know if it is normal or even if it matters though, I'll keep using
the "this" keyword in this part of the condition for now.
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
7:18 a.m.
Also, with the "before" keyword in place of "after", the problem is
different, it considers all events as the start event.
rule "Detect it"
when
$start : RawEvent()
not $re : RawEvent(eval($re != $start), id == $start.id, this
before[0ms, 90s] $start)
then
System.out.println("detected: " + $start);
end
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
7:37 a.m.
Hi again,
I tried using 5.3 and I noticed the "before" keyword has the same problem as
5.4 :
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, this before[0ms, 90s]
$start)
then
System.out.println("detected: " + $start);
end
Gives a null pointer exception.
Is this really a bug ? Or am i misusing the "before" keyword ?
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
Sent from the Drools: User forum mailing list archive at Nabble.com.
10:42 a.m.
With 5.3.0, I find that there's no difference between these two forms.
rule "Detect it"
when
$start : RawEvent()
not RawEvent( this != $start,
id == $start.id,
## $start after[0ms, 90s] this )
this before[0ms, 90s] $start )
then
System.out.println("detected:" + $start);
end
-W
On 3 September 2012 14:37, Rob21 <rob.perrez(a)gmail.com> wrote:
Hi again,
I tried using 5.3 and I noticed the "before" keyword has the same problem
as
5.4 :
rule "Detect it"
when
$start : RawEvent()
not RawEvent(this != $start, id == $start.id, this before[0ms, 90s]
$start)
then
System.out.println("detected: " + $start);
end
Gives a null pointer exception.
Is this really a bug ? Or am i misusing the "before" keyword ?
--
View this message in context:
http://drools.46999.n3.nabble.com/before-after-behavior-for-negative-patt...
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
4250
days inactive
4252
days old
11 comments
2 participants
participants (2)
-
Rob21 -
Wolfgang Laun