Re: [rules-users] determining what field changed

In general do you have any suggestions for getting around multiple update detects being made? In other words I have a rule to detect if my DataObject has fieldA > X and if so, then modify fieldB. Then another rule that checks if fieldB = value and then updates field C. So something like: rule "identify modified DATAOBJECT" when do : DataObject(FieldA > x); then modify(do.fieldB = value) end So something like: rule "identify modified DATAOBJECT" when do : DataObject(fieldB == value); then modify(do.fieldC = someValue) end rule "identify modified DATAOBJECT" when do : DataObject(fieldC == someValToMatchForC); then modify(do.fieldA = someNewValueForA) end But of course this leads to some sort of cross product execution as each one is run all over again after another makes a change and ramps out of control One question. I have been using modify() and I notice you had explicit: do.setN00b(false); update(do); What is the difference and will that matter to my scenario? Thanks, Chris -----Original Message----- From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Greg Barton Sent: Wednesday, September 16, 2009 10:14 AM To: Rules Users List Subject: Re: [rules-users] determining what field changed Not really. Without modifying the objects you could do this outside the rules using a WorkingMemoryEventListener, which has objectInserted() and objectUpdated() methods. Event if you can modify the DataObject to add an "I'm a new object" flag, and set it immediately on insertion, then you could do differentiate with two rules: rule "identify new DATAOBJECT" when do : DataObject(n00b == true); then do.setN00b(false); update(do); System.err.println("DataObject assertion detected: " + do); end rule "identify modified DATAOBJECT" when do : DataObject(n00b == false); then System.err.println("DataObject modification detected: " + do); end Which would still catch a spurious modification notification on assert! At that point the semi-risky alternative is to not call update() in the "identify new DATAOBJECT" rule action. That way, the "identify modified DATAOBJECT" is not immediately activated, but it would be activated the next time update() is called on the object. I say semi-risky because it's not often a good idea to put a WM object into a state that the rete is not aware of, but in this case it would do what you want. --- On Wed, 9/16/09, Chris Richmond <crichmond(a)referentia.com&gt; wrote: _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users