9:51 a.m.
Hello,
First of all i have to appologise for this horrible joke, but i think you
understand that i'm interrested in shadow facts.
I read that theses lazy facts proxy were introduced in Drools 4 in order to
make the rule engine/working memory consistent. I have some questions about
the general implentation of shadow facts in drools 5.4 :
- Did the shadow facts been implemented since version 4, or was it earlyier?
- Are the shadow facts still implemented in drools 5.4 final?
- How the mapping between the fact fields and the proxy fields is
implemented ? By this i want to know all fields are copy when the fact is
insert into the session.
- Is there the possibility to disable shadow facts ?
You have to know that we get the knowledgeBase through a knowledgeAgent
which
Thank you.
--
View this message in context: http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
10:07 a.m.
On 24/05/2013, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Hello,
First of all i have to appologise for this horrible joke, but i think you
understand that i'm interrested in shadow facts.
OK, but you better be careful - I was inclined to can this mail due to
its subject and the preview of the first line...
I read that theses lazy facts proxy were introduced in Drools 4 in
order to
make the rule engine/working memory consistent. I have some questions about
the general implentation of shadow facts in drools 5.4 :
- Did the shadow facts been implemented since version 4, or was it earlyier?
- Are the shadow facts still implemented in drools 5.4 final?
- How the mapping between the fact fields and the proxy fields is
implemented ? By this i want to know all fields are copy when the fact is
insert into the session.
- Is there the possibility to disable shadow facts ?
There is no such p(a)roxysm in Drools 5. Objects are inserted into
Working Memory to become facts, and that's all there is to it. Period.
-W
You have to know that we get the knowledgeBase through a knowledgeAgent
which
Thank you.
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948.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
10:15 a.m.
On 24 May 2013, at 08:51, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Hello,
First of all i have to appologise for this horrible joke, but i think you
understand that i'm interrested in shadow facts.
I read that theses lazy facts proxy were introduced in Drools 4 in order to
make the rule engine/working memory consistent. I have some questions about
the general implentation of shadow facts in drools 5.4 :
They did exist in 4, we
had a similar engine to Jess then.
http://blog.athico.com/2008/02/shadow-facts-what-you-always-wanted-to.html
They are no longer needed in 5, due to changes in our rete algorithm. See Assymetrical
Rete algorithm:
http://blog.athico.com/2008/07/drools-50-m1-new-and-noteworthy.html
http://blog.athico.com/2008/10/symmetrical-and-asymmetrical-rete.html
http://blog.athico.com/2010/01/rete-and-true-modify.html
- Did the shadow facts been implemented since version 4, or was it
earlyier?
- Are the shadow facts still implemented in drools 5.4 final?
- How the mapping between the fact fields and the proxy fields is
implemented ? By this i want to know all fields are copy when the fact is
insert into the session.
- Is there the possibility to disable shadow facts ?
Don't need to, they are
gone.
You have to know that we get the knowledgeBase through a knowledgeAgent
which
Thank you.
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948.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
10:50 a.m.
Ok i understand that the need to have proxies no longer exits due to new
asymetric rete algorithm. I but i'm a little bit desapointed. It seem that
when a fact is introduced into the working memory, all it's fields are
evaluate.
I guess i have to explain what is my problem. I have business objects which
are directly used into the working memory. Theses BO are not entirely POJO,
some of them perform SQL queries. We used interfaces to produced a POJO jar,
which have been sent to knowledge base.
Previously we used drools in version 3, but now we use version 5.4 and we
notice a significant performance reduction.
It seems that all fields of the BO are readed even if they aren't used. Our
rules are decision table in spreadsheet format. Each row's start with a
condition on a fields which doesn't make a SQL query, so normaly there isn't
a lot of peformed SQL queries.
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023951.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
4:02 p.m.
On 24 May 2013, at 09:50, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Ok i understand that the need to have proxies no longer exits due to
new
asymetric rete algorithm. I but i'm a little bit desapointed. It seem that
when a fact is introduced into the working memory, all it's fields are
evaluate.
That is not true.
Only the fields used are the ones evaluated. There is no pre-processing, they are only
touched, when the constraint itself is evaluated.
I guess i have to explain what is my problem. I have business objects
which
are directly used into the working memory. Theses BO are not entirely POJO,
some of them perform SQL queries. We used interfaces to produced a POJO jar,
which have been sent to knowledge base.
Previously we used drools in version 3, but now we use version 5.4 and we
notice a significant performance reduction.
It seems that all fields of the BO are readed even if they aren't used.
This
would not be a problem on the Drools side, something else must be forcing all field sot be
read - if you are not writing constraints to read them.
Our
rules are decision table in spreadsheet format. Each row's start with a
condition on a fields which doesn't make a SQL query, so normaly there isn't
a lot of peformed SQL queries.
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023951.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
4:35 p.m.
Ok, thank you for this explaination. I know that the salience can build a
priority between rules into a single excel file. But it there a priority
criteria between column? For exemple from left column to right.
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023953.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
8:47 p.m.
On 24 May 2013, at 15:35, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Ok, thank you for this explaination. I know that the salience can
build a
priority between rules into a single excel file. But it there a priority
criteria between column? For exemple from left column to right.
no
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023953.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
9:55 a.m.
Is there a way to force a rule evaluation from left to right? For example if
i have :
eval(cond1) | eval(cond2)
I wan't that cond2 is evaluate only if cond1 is true. Is this possible to
have this behaviour?
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023980.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
9:58 a.m.
A single eval(), with terms separated by "||" with the same semantics
as in Java.
-W
On 27/05/2013, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Is there a way to force a rule evaluation from left to right? For
example
if
i have :
eval(cond1) | eval(cond2)
I wan't that cond2 is evaluate only if cond1 is true. Is this possible to
have this behaviour?
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023980.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
10 a.m.
...separated by "||" and "&&", as required, with...
On 27/05/2013, Wolfgang Laun <wolfgang.laun(a)gmail.com> wrote:
A single eval(), with terms separated by "||" with the same
semantics
as in Java.
-W
On 27/05/2013, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
> Is there a way to force a rule evaluation from left to right? For example
> if
> i have :
> eval(cond1) | eval(cond2)
> I wan't that cond2 is evaluate only if cond1 is true. Is this possible to
> have this behaviour?
>
>
>
> --
> View this message in context:
> http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023980.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
>
4:17 p.m.
it already works that way, eval cond2 will only be evaluated after cond1.
Mark
On 27 May 2013, at 08:55, ronan.quintin <ronan.quintin(a)gmail.com> wrote:
Is there a way to force a rule evaluation from left to right? For
example if
i have :
eval(cond1) | eval(cond2)
I wan't that cond2 is evaluate only if cond1 is true. Is this possible to
have this behaviour?
--
View this message in context:
http://drools.46999.n3.nabble.com/Shadow-fax-tp4023948p4023980.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
4262
days inactive
4265
days old
10 comments
3 participants
participants (3)
-
Mark Proctor -
ronan.quintin -
Wolfgang Laun