JESS: Type precision causes duplicate facts

Hello,

As part of testing the problem from my previous posting, I noticed a strange duplication of facts when trying to create a fact with a slot of FLOAT type. In one case I am specifying a decimal and the other a whole number, but the number is otherwise the same.  I would think only one fact would be created, but instead two are created but with different float representations.  The values are equal according to the = operator; however the eq operator returns false.

 

Is this behavior to be expected and intended by design?   I would think the slot with type Float would automatically convert all raw values to float and so only create one fact in the example below.  

 

(deftemplate x "blah"

            (slot num (type FLOAT)))

(defrule dupFact

    ?x1 <- (x (num ?num1))

    ?x2 <- (x (num ?num2&:(and (<> ?x1 ?x2) (= ?num1 ?num2))))

    =>

            (printout t "dup facts with values " ?num1 " " ?num2 crlf)

    )

 

(reset)

 

(assert (x (num 1157650583.0)))

(assert (x (num 1157650583)))

 

(run)

 

(facts)

 

----

Jess, the Rule Engine for the Java Platform

Copyright (C) 2008 Sandia Corporation

Jess Version 7.1p2 11/5/2008

 

MAIN::dupFact: +1+1+2+t

 ==> Focus MAIN

 ==> f-0 (MAIN::initial-fact)

 ==> f-1 (MAIN::x (num 1.157650583E9))

 ==> f-2 (MAIN::x (num 1157650583))

==> Activation: MAIN::dupFact :  f-2, f-1

==> Activation: MAIN::dupFact :  f-1, f-2

FIRE 1 MAIN::dupFact f-1, f-2

dup facts with values 1.157650583E9 1157650583

FIRE 2 MAIN::dupFact f-2, f-1

dup facts with values 1157650583 1.157650583E9

 <== Focus MAIN

f-0   (MAIN::initial-fact)

f-1   (MAIN::x (num 1.157650583E9))

f-2   (MAIN::x (num 1157650583))

For a total of 3 facts in module MAIN.

 

Thanks,

Brian

On Jun 8, 2009, at 9:18 PM, Brian Rogosky wrote:

> Hello,
> As part of testing the problem from my previous posting, I noticed a  
> strange duplication of facts when trying to create a fact with a  
> slot of FLOAT type. In one case I am specifying a decimal and the  
> other a whole number, but the number is otherwise the same.  I would  
> think only one fact would be created, but instead two are created  
> but with different float representations.  The values are equal  
> according to the = operator; however the eq operator returns false.

The whole number is an RU.INTEGER, and the float is an RU.FLOAT.  
Comparing with = returns true because they have the same value, but eq  
returns false because they're different data types.

Variables and slots are untyped in Jess; only values have types. The  
"type" slot qualifier is included only for backwards compatibility  
with CLIPS; in Jess, as in CLIPS, it doesn't actually prevent a value  
of another type from being added to the slot. It's more for  
documentation than anything else. It's definitely not like a variable  
declaration in Java.

>
> Is this behavior to be expected and intended by design?

Yes.


>  I would think the slot with type Float would automatically convert  
> all raw values to float and so only create one fact in the example  
> below.

Nope. 1 and 1.0 are two different values. When it matters, use "=" for  
comparison.


---------------------------------------------------------
Ernest Friedman-Hill
Informatics & Decision Sciences          Phone: (925) 294-2154
Sandia National Labs
PO Box 969, MS 9012                            ejfried@...
Livermore, CA 94550                             http://www.jessrules.com





--------------------------------------------------------------------
To unsubscribe, send the words 'unsubscribe jess-users you@...'
in the BODY of a message to majordomo@..., NOT to the list
(use your own address!) List problems? Notify owner-jess-users@....
--------------------------------------------------------------------