The correct way to ascertain that something hasn't happened for
a certain interval must be based on "staking out" this interval. One
way is by means of two distinct events:
when
$r1: Reading( reading >90 )
$r2: Reading( reading > 90, this after[30s] $r1 )
not Reading( reading <= 90, this after $r1 && before $r2 )
then
Clearly, this will only work if Readings arrive regularly, and you may have a
certain amount of delay at the end.
Another way would be to use a sliding window over time:
when
not Reading( reading <= 90 ) over window:time( 30s )
then
Here's a snag, too: If there are no readings at all, you'll also get an alert.
And then you'll have to think about the seconds following the 30th
second: should another alert be emitted repeatedly...
Most of the time, it's not an issue in CEP ;-)
-W
On 27/02/2014, lin.lin <lin.lin(a)comcast.net> wrote:
Hi,
I tried to create a rule that gives an alert when a machine reading is high
for 30 seconds. I modeled the machine reading input as events. The
following is the rule:
rule "rule1"
dialect "mvel"
when
$rd1 : mReading( reading >90 ) from entry-point ReadingStream
not ($rd2 : mReading (reading<=90 , this before [0s, 30s] $rd1 )
from entry-point ReadingStream )
then
System.out.print ("alert: Reading high for 30 seconds \n");
end
This does not detect detect the bad reading. However, If I changed the
rule
to be following (change second condition from "not" to "exists"), it
detects
all readings that are good.
rule "rule1"
dialect "mvel"
when
$rd1 : mReading( reading >90 ) from entry-point ReadingStream
exists ($rd2 : mReading (reading<=90 , this before [0s, 30s] $rd1 )
from entry-point ReadingStream )
then
System.out.print ("alert: Reading high for 30 seconds \n");
End
Is this an issue in CEP or did I do something wrong?
Thanks a lot!
Lily
--
View this message in context:
http://drools.46999.n3.nabble.com/sliding-window-question-tp4028367.html
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