Absence Pattern question
by wendy
Hi,
I'm having trouble writing an absence pattern. What I'm trying to do is
detect when there is an A followed by no Bs for 3+minutes. I don't care if
there is more than one A. What I'm running into is that when I try to use
'over window:time' the time within drools is the end time of the window. So
this means I need to write the no Bs for 3+ min first:
not( $b: B() over window:time(3m))
then try to find the A before it:
$now: Long() from RuleUtilityFunctions.getSessionClockTime()
$a: A( this before [3m] $now)
then I should have to check to make sure that I don't have any Bs between
$a's time and the start of the no B window:
not( B( time >= $a.time,
time <= $now))
This is not working. I think that it has something to do with my function
to get the session clock time (RuleUtilityFunctions.getSessionClockTime())
and how things get evaluated within the Rete engine. Because it does not
seem like $now is getting re-evaluated on future calls that pass the $b
condition. If I replace $now with the call to getSessionClockTime()
everything just seems to get weird.
I've tried to write the rule forward too. Look for A followed by no B but
that does not seem to work because A is matched at the current time and the
rule triggers because there is no B because the future B data has not be
inserted into working memory yet.
What is the right way to write this rule? Is there a way to get the start
and end time of the time window that met the over window:time() condition?
Thank you,
Wendy
--
View this message in context: http://drools.46999.n3.nabble.com/Absence-Pattern-question-tp3140377p3140...
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
Drools Clustering
by dan.dobrescu
Hi,
I've searched the forum, but I couldn't find nothing useful about my
question..
I've got a little problem.. So, I've implemented in Drools some rules,
mostly triggered by two external events... EVENTA, and EVENTB. My
applications is clustered, so it's possible that EVENTA is inserted from
NODE1 and EVENTB from NODE2. But in this case, I'll have different facts in
both sessions, and it's possible that my rules don't fire. Has drools
implemented clustering for statefull sessions? or how should I do, because
if both session are containing the same facts, the rules will be triggered
twice... I'd like a behavior similar to Quartz..
--
View this message in context: http://drools.46999.n3.nabble.com/Drools-Clustering-tp3140305p3140305.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
org.hibernate.exception.LockAcquisitionException: could not delete:
by Jose
Hi All,
When i tried to execute the rule in multi threaded environment, i got
the below mentioned Error logs...
I just tried to execute the rule flow with 2 thread, always 1 got executed
successfully and one fails with the following error log.
ERROR LOG:
Hibernate: delete from ProcessInstanceInfo where InstanceId=? and OPTLOCK=?
Hibernate: delete from ProcessInstanceInfo where InstanceId=? and OPTLOCK=?
2011/07/04 11:20:06.475{UTC} [JDBCExceptionReporter :
JDBCExceptionReporter.java: logExceptions: 100] [RMI TCP
Connection(6)-10.50.53.32]: WARN: : SQL Error: 60, SQLState: 61000
2011/07/04 11:20:06.475{UTC} [JDBCExceptionReporter :
JDBCExceptionReporter.java: logExceptions: 101] [RMI TCP
Connection(6)-10.50.53.32]: ERROR: : ORA-00060: deadlock detected while
waiting for resource
2011/07/04 11:20:06.475{UTC} [AbstractFlushingEventListener :
AbstractFlushingEventListener.java: performExecutions: 324] [RMI TCP
Connection(6)-10.50.53.32]: ERROR: : Could not synchronize database state
with session
org.hibernate.exception.LockAcquisitionException: could not delete:
[org.drools.persistence.processinstance.ProcessInstanceInfo#8130]
at
org.hibernate.exception.SQLStateConverter.convert(SQLStateConverter.java:112)......
Thanks,
Jose.
--
View this message in context: http://drools.46999.n3.nabble.com/org-hibernate-exception-LockAcquisition...
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
check whether a key is present in global Map, i.e. mimicking containsKey() behavior
by esstrata
Hello,
I'm using Drools 5.1.1 in a JavaSE 6 project.
I can't get running a rule where a value obtained from WM is check for
existence as a key of a global Map.
I have a map of the type ProductTextCode -> Product, where keys are String
and values Product
/global Map productMap/
Then in a rule a I want a condition where each String in WM is tested
against map keyset.
/
rule "Product ID not in products map"
dialect "mvel"
when
$pid : String()
=> Here goes the condition <=
then
System.out.println("Invalid product id: " + $pid)
retract ($pid)
end
/
So far I've tried with productMap.containsKey($pid) but a get a MVEL error
as it takes 'containsKey' as a property.
Please help. Thank you.
--
View this message in context: http://drools.46999.n3.nabble.com/check-whether-a-key-is-present-in-globa...
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
Guvnor errors stay forever
by Justin Case
Hi all,
Here's a funny thing I experience:
I declare a function and do the mistake of verifying it before editing anything,
while it's still a template (so it has chevrons). The verify obviously fails -
but with a HTTP error 500! Server crashed?? Indeed I can see in the console:
at
org.apache.tomcat.util.net.JIoEndpoint$Worker.run(JIoEndpoint.java:951)
[:6.0.0.Final]
at java.lang.Thread.run(Thread.java:619) [:1.6.0_12]
Caused by: java.lang.RuntimeException: Verifier Errors:
[ERR 102] Line 1:9 mismatched input '<' in function
[ERR 102] Line 1:36 mismatched input '<' in function "calculateAge"
[ERR 102] Line 1:46 mismatched input '>' in function "calculateAge"
Parser returned a null Package
I see quite a few JIRAs about incomplete exception handling - GUVNOR-741,
GUVNOR-690, GUVNOR-536, should I add the above to any of them?
Whatever, an error is an error, but here comes the real problem:
After the above I come to realize that each analysis of a package keeps adding
the results of the previous ones, so the Errors/Warnings list just grows and
grows. An old error will always come again even though the item it refers to is
not even existing anymore (so I cannot analyse/verify anything reliably). I
renamed the above function, I edited it so it doesn't have the chevrons, its
code can be validated fine, but nevertheless Guvnor keeps throwing the same
error 500 and the JBoss console keeps complaining about the oooold function
calculateAge and its chevrons each time I verify or analyse. And it does that by
repeating the message for each retry once more. Either the error list is not
cleaned after the verify, or some artifacts live too long. Logout and relogon
keeps the list, but restarting JBoss solves the problem.
.
I really have the feeling GUVNOR-1499 has the same underlying problem...
For the record: I'm on 5.2.0.Final on JBoss 6, Java 6 on Windows XP.
Thank you very much,
JC
15 years
Retrieving Enumerated List from Drools
by lansyj
Hi,
We have a custom CMS which we use to manage content like banner images, and
other personalization content.
We would be using Drools to define user segments and would be associating a
unique code and short name to it - code for the system and short name for
the business users.
Now, is there a way for our custom CMS to be able to retrieve the enumerated
list (basically all possible values of that data element) such that it can
be used for mapping content in the CMS?
Any tip would be helpful.
Best Regards
-lj
--
View this message in context: http://drools.46999.n3.nabble.com/Retrieving-Enumerated-List-from-Drools-...
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
Using Enumeration for long lists
by lansyj
Hi,
We are planning to use Drools for a Blacklist validation logic which
requires us to manage very long list (could be tens of thousands of IDs).
Can we use the native Drools enumeration for this? Would this be an optimal
implementation?
The other option we were thinking of was to have a list managed separately
in our core application and have in lucene indexed such that Drools rule
checks against the index. But, as you can see, it creates a lot of
development as well as maintenance overhead.
Thanks in advance for any tips on this topic.
Best Regard
-lj
--
View this message in context: http://drools.46999.n3.nabble.com/Using-Enumeration-for-long-lists-tp3140...
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
delegate a task
by damon
Hi,
I call MinaTaskClient's delegate() method to delegate a task to another
user, I found the state of the task is translate to '*Ready*', instead of
'*Reserved*'.
But from the 'Web Services Human Task Version 1.0' specification, The state
will be translate to 'Reserved' when do delegate -- "A task can be delegated
when it is in an active state (Ready, Reserved, InProgress), and transitions
the task into the Reserved state"
Is there any wrong when I call delegate() method? Or any precondition when
call delegate() method, then the state will be translate to 'Reserved',
instead of 'Ready'?
Thanks
Damon
--
View this message in context: http://drools.46999.n3.nabble.com/delegate-a-task-tp3137901p3137901.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
15 years
