Hello. I'm looking for the better way to write a rule that collects data only when they have been all processed. For example, two input objects classes as facts : *Data* and *Category*. I declare a local type to memorize the result of the evaluation for each pair of fact. Some rule evaluates the Cartesian product of all *Data* by all *Category* : Now I want to collect all the results, but only when the previous rule has been fired for all the data. Something like : Any advice ? Thanks in advance for your help. Regards. -- View this message in context: http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... 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

Hello. Thanks for this advice. I've tried it with a small test case, and it actually works fine : the collecting rule seems to be fired at the end. Nonetheless, I'm not confident at all with salience, since in my real use case I have many rules before and after that, and I worry that salience might move the problem elsewhere, and values must be set to all of them to ensure they're all executed according to the expected order. At this moment, the rules before the collecting rule also produce facts that are expected by the rules that follow it, and, I suppose, it's for that reason that all the ending part of the job is executed after the correct collection of the intermediary results. Another alternative ? Thanks again. Regards Ephemeris Lappis Le 21/04/2014 13:11, laune [via Drools] a écrit : > This is one of the (rare) cases where I'd advocate salience. > > rule "Collect results when complete" > salience -999999 > when > $all : List from collect(Result()) > then > ...do something with $all... > end > > -W > > > On 21/04/2014, Ephemeris Lappis <[hidden email] > </user/SendEmail.jtp?type=node&node=4029298&i=0>> wrote: > > > Hello. > > > > I'm looking for the better way to write a rule that collects data > only when > > they have been all processed. > > > > For example, two input objects classes as facts : *Data* and > *Category*. > > > > I declare a local type to memorize the result of the evaluation for > each > > pair of fact. > > > > > > > > Some rule evaluates the Cartesian product of all *Data* by all > *Category* : > > > > > > > > Now I want to collect all the results, but only when the previous > rule has > > been fired for all the data. > > Something like : > > > > > > > > Any advice ? > > > > Thanks in advance for your help. > > > > Regards. > > > > > > > > -- > > View this message in context: > > > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > > Sent from the Drools: User forum mailing list archive at Nabble.com. > > _______________________________________________ > > rules-users mailing list > > [hidden email] </user/SendEmail.jtp?type=node&node=4029298&i=1> > > https://lists.jboss.org/mailman/listinfo/rules-users > > > _______________________________________________ > rules-users mailing list > [hidden email] </user/SendEmail.jtp?type=node&node=4029298&i=2> > https://lists.jboss.org/mailman/listinfo/rules-users > > > ------------------------------------------------------------------------ > If you reply to this email, your message will be added to the > discussion below: > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > > > To unsubscribe from Collect all occurrences of resulting data., click > here > <http://drools.46999.n3.nabble.com/template/NamlServlet.jtp?macro=unsubscr...;. > NAML > <http://drools.46999.n3.nabble.com/template/NamlServlet.jtp?macro=macro_vi... > > -- View this message in context: http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... Sent from the Drools: User forum mailing list archive at Nabble.com.

Hello again. It seems clear that too many salience levels would make it all worse ! My ruleset executes over a simple list of initial facts, and proceeds more or less as you say in three phases : 1) Analise initial facts using several deduced facts to make intermediary conclusions implemented as locally declared types. 2) Consolidate the scattered produced facts to build seven main business rules : here is where I need to gather all the facts into a unique collection. Depending on each of the seven rules, the result may use a simple input fact, or, at the opposite, different collections of temporary facts... 3) At the end, for each input fact, generate a global decision using the seven rules, and output it into a global map. For that, I think that it may be rather reasonable [:)] setting 3 salience levels representing the three main steps of the rule system. I have had a look to the activation groups, but I don't think it could match this current need. Moreover this would be more intrusive since such a way the Java code should knows the layered structure of the rules. At this time the application code just puts the initial facts and extract the resulting map to produce a JSON like object to be written into a Mongo base... Thanks a lot for your help. Any new advice is welcome ! Best regards. Ephemeris Lappis Le 21/04/2014 17:48, laune [via Drools] a écrit : > I recommend at most 3 levels of salience, and only if you have a > very limited scenario of insert / fire. > > Alternatively, you can use agenda groups. > > If I understand you correctly, you have some rules that execute > "Before" and others that should execute "After". In your .drl, > > activation-group "Before" > rule ... when ... then ... end ... rule ... when ... then ... end > activation-group "After" > rule ... when ... then ... end ... rule ... when ... then ... end > > You can set (the group "Before" before you insert facts, call > fireAllRules, set "After" and fire again; then start over again. - See > the Expert manual for some more about activation groups. > > Setting the agenda group from Java takes a little digging in the API doc: > session.getAgenda().getAgendaGroup( "Before" ).setFocus(); > > -W > > > > On 21/04/2014, Ephemeris Lappis <[hidden email] > </user/SendEmail.jtp?type=node&node=4029304&i=0>> wrote: > > > Hello. > > > > Thanks for this advice. > > > > I've tried it with a small test case, and it actually works fine : the > > collecting rule seems to be fired at the end. > > > > Nonetheless, I'm not confident at all with salience, since in my real > > use case I have many rules before and after that, and I worry that > > salience might move the problem elsewhere, and values must be set to > all > > of them to ensure they're all executed according to the expected order. > > > > At this moment, the rules before the collecting rule also produce facts > > that are expected by the rules that follow it, and, I suppose, it's for > > that reason that all the ending part of the job is executed after the > > correct collection of the intermediary results. > > > > Another alternative ? > > > > Thanks again. > > > > Regards > > > > > > > > Ephemeris Lappis > > > > Le 21/04/2014 13:11, laune [via Drools] a écrit : > >> This is one of the (rare) cases where I'd advocate salience. > >> > >> rule "Collect results when complete" > >> salience -999999 > >> when > >> $all : List from collect(Result()) > >> then > >> ...do something with $all... > >> end > >> > >> -W > >> > >> > >> On 21/04/2014, Ephemeris Lappis <[hidden email] > >> </user/SendEmail.jtp?type=node&node=4029298&i=0>> wrote: > >> > >> > Hello. > >> > > >> > I'm looking for the better way to write a rule that collects data > >> only when > >> > they have been all processed. > >> > > >> > For example, two input objects classes as facts : *Data* and > >> *Category*. > >> > > >> > I declare a local type to memorize the result of the evaluation for > >> each > >> > pair of fact. > >> > > >> > > >> > > >> > Some rule evaluates the Cartesian product of all *Data* by all > >> *Category* : > >> > > >> > > >> > > >> > Now I want to collect all the results, but only when the previous > >> rule has > >> > been fired for all the data. > >> > Something like : > >> > > >> > > >> > > >> > Any advice ? > >> > > >> > Thanks in advance for your help. > >> > > >> > Regards. > >> > > >> > > >> > > >> > -- > >> > View this message in context: > >> > > >> > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > >> > Sent from the Drools: User forum mailing list archive at Nabble.com. > >> > _______________________________________________ > >> > rules-users mailing list > >> > [hidden email] </user/SendEmail.jtp?type=node&node=4029298&i=1> > >> > https://lists.jboss.org/mailman/listinfo/rules-users > >> > > >> _______________________________________________ > >> rules-users mailing list > >> [hidden email] </user/SendEmail.jtp?type=node&node=4029298&i=2> > >> https://lists.jboss.org/mailman/listinfo/rules-users > >> > >> > >> > ------------------------------------------------------------------------ > >> If you reply to this email, your message will be added to the > >> discussion below: > >> > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > >> > >> > >> To unsubscribe from Collect all occurrences of resulting data., click > >> here > >> < > >> NAML > >> > <http://drools.46999.n3.nabble.com/template/NamlServlet.jtp?macro=macro_vi... > > > >> > >> > > > > > > > > > > > > -- > > View this message in context: > > > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > > Sent from the Drools: User forum mailing list archive at Nabble.com. > > _______________________________________________ > rules-users mailing list > [hidden email] </user/SendEmail.jtp?type=node&node=4029304&i=1> > https://lists.jboss.org/mailman/listinfo/rules-users > > > ------------------------------------------------------------------------ > If you reply to this email, your message will be added to the > discussion below: > http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... > > > To unsubscribe from Collect all occurrences of resulting data., click > here > <http://drools.46999.n3.nabble.com/template/NamlServlet.jtp?macro=unsubscr...;. > NAML > <http://drools.46999.n3.nabble.com/template/NamlServlet.jtp?macro=macro_vi... > > -- View this message in context: http://drools.46999.n3.nabble.com/Collect-all-occurrences-of-resulting-da... Sent from the Drools: User forum mailing list archive at Nabble.com.