Just looking at your rules, I believe it is a problem in understanding
how sliding windows work in Drools.
In Drools, sliding windows are used to filter events based on timestamp,
but they do not prevent the detection of an event. I.e., although your event
might have arrived late (for whatever reason, including network latency), it
still happened, and since you are using no other filters, your rule means
you wanted to detect that event. So, an activation is created for that rule,
even if the event will be expired right after the activation is created.
Please note that this is not the only semantics possible for Sliding
Windows found in the literature, but was the one that best matched our
requirements and was therefore adopted for the Drools implementation.
I want to add more semantics and make them configurable in the future,
but it is not there yet.
Edson
2010/2/25 Venkat K <nabble-venkatp(a)sneakemail.com>
Firstly, We very much appreciate Drools and it's developers. Thanks in
advance for the community for any help . To add to my colleague's points,
1. STREAM mode is turned on.
2. We have tried the before operator and it works as expected.
3. We have also tried the auto timestamp as opposed to using a property of
the event itself for timestamping. To test sliding time window, we put
sleeps between insert of objects to get gaps between the inserts.
It would be great if anyone can share some light on what we may be doing
wrong.
Thanks and Regards,
Venkat.
--
View this message in context:
http://n3.nabble.com/Re-Using-Sliding-Time-Window-usage-in-Drools-Fusion-...
Sent from the Drools - User mailing list archive at
Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
Edson Tirelli
JBoss Drools Core Development
JBoss by Red Hat @
www.jboss.com