4:47 a.m.
Hi All
I am new to drools and I am trying to see if we can use it. I have a simple
file
public class Data {
private int id =0;
public Data(int id) {
this.id = id;
}
public int getId() {
return id;
}
}
rule "Unique data"
when
data : Data()
old : Data(this != data, this.id == data.id)
then
log.log("Following data are not unique: " + data.getId() + " and
" +
old.getId());
end
When I try to load 30,000 of data in memory it takes long time ( around 12
hours ) Can anybody suggests a better way of doing it.
Thanks
Gaurav
--
View this message in context:
http://www.nabble.com/simple-rule-takes-long-time-tp18192098p18192098.html
Sent from the drools - user mailing list archive at Nabble.com.
9:47 a.m.
1) Finding all the duplicate pairs is causing combinatorial explosion.
With 30000 objects the rule will compare every object against every
other object which requires running the run about 450 MILLION times. If
you must discover every pair that is a duplicate, then I think you have
a hard problem. However, if you can stop considering an object once you
see it is a duplicate of at least one other datum, then you can add a
field to your object to record that it is already a duplicate and remove
it from further consideration, something like the following:
public class Data {
private int id =0;
private boolean isDuplicated = false;
public Data(int id) {
this.id = id;
}
public int getId() {
return id;
}
public boolean getIsDuplicated() {
return isDuplicated;
}
public void setIsDuplicated( boolean isDuplicated ) {
this.isDuplicated = isDuplicated;
}
}
rule "Unique data"
when
data : Data()
old : Data( isDuplicated == false, this != data, this.id == data.id)
then
log.log("Following data are not unique: " + data.getId() + " and
" + old.getId());
$data.setIsDuplicated( true );
$old.setIsDuplicated( true );
end
--------------------
2) I'm pretty sure the test, this != $data, is redundant since it's
understood that the condition is relating two *different* objects of the
same class. Tagging the id fields will make rule a little more efficient
and easier to read. There's no need to use 'this.id', it's understood
that 'id' is the id field of old.
rule "Unique data"
when
$data1: Data( $id1: id )
$data2: Data( isDuplicated == false, id == $id1 )
then
log.log( "Duplicated id: " + $id1 );
$data1.setIsDuplicated( true );
$data2.setIsDuplicated( true );
end
( I like to use dollar signs at the front of tag names because it makes
it easier to distinguish what is what. You don't have to. )
ygaurav wrote:
Hi All
I am new to drools and I am trying to see if we can use it. I have a simple
file
public class Data {
private int id =0;
public Data(int id) {
this.id = id;
}
public int getId() {
return id;
}
}
rule "Unique data"
when
data : Data()
old : Data(this != data, this.id == data.id)
then
log.log("Following data are not unique: " + data.getId() + " and
" +
old.getId());
end
When I try to load 30,000 of data in memory it takes long time ( around 12
hours ) Can anybody suggests a better way of doing it.
Thanks
Gaurav
10:34 a.m.
I forgot the updates and added a couple of $'s (see below).
Scott Reed wrote:
1) Finding all the duplicate pairs is causing combinatorial
explosion.
With 30000 objects the rule will compare every object against every
other object which requires running the run about 450 MILLION times.
If you must discover every pair that is a duplicate, then I think you
have a hard problem. However, if you can stop considering an object
once you see it is a duplicate of at least one other datum, then you
can add a field to your object to record that it is already a
duplicate and remove it from further consideration, something like the
following:
public class Data {
private int id =0;
private boolean isDuplicated = false;
public Data(int id) {
this.id = id;
}
public int getId() {
return id;
}
public boolean getIsDuplicated() {
return isDuplicated;
}
public void setIsDuplicated( boolean isDuplicated ) {
this.isDuplicated = isDuplicated;
}
}
rule "Unique data"
when
data : Data()
old : Data( isDuplicated == false, this != data, this.id ==
data.id)
then
log.log("Following data are not unique: " + data.getId() + "
and " + old.getId());
$data.setIsDuplicated( true );
$old.setIsDuplicated( true );
The previous two lines should be:
data.setIsDuplicated( true );
update(data);
old.setIsDuplicated( true );
update(old);
end
--------------------
2) I'm pretty sure the test, this != $data, is redundant since it's
understood that the condition is relating two *different* objects of
the same class. Tagging the id fields will make rule a little more
efficient and easier to read. There's no need to use 'this.id', it's
understood that 'id' is the id field of old.
rule "Unique data"
when
$data1: Data( $id1: id )
$data2: Data( isDuplicated == false, id == $id1 )
then
log.log( "Duplicated id: " + $id1 );
$data1.setIsDuplicated( true );
$data2.setIsDuplicated( true );
end
( I like to use dollar signs at the front of tag names because it
makes it easier to distinguish what is what. You don't have to. )
ygaurav wrote:
> Hi All
> I am new to drools and I am trying to see if we can use it. I have a
> simple
> file
>
> public class Data {
> private int id =0;
> public Data(int id) {
> this.id = id;
> }
> public int getId() {
> return id;
> }
> }
>
> rule "Unique data"
> when
> data : Data()
> old : Data(this != data, this.id == data.id)
> then log.log("Following data are not unique: " +
> data.getId() + " and " +
> old.getId());
> end
>
>
> When I try to load 30,000 of data in memory it takes long time (
> around 12
> hours ) Can anybody suggests a better way of doing it.
>
> Thanks
> Gaurav
>
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
4:48 a.m.
Hi All
I have few changes to my data class as suggested above and rule and it seems
to work. But that brings me to another question. Here is my new data class
Data class
public class Data {
private int id =0;
public boolean isCheck = false;
public String name = "";
public Data(int id) {
this.id = id;
name = "name"+(int)(this.hashCode());
System.out.println("id is "+id+" name is "+name);
}
public int getId() {
return id;
}
public void setIsCheck() {
isCheck = true;
}
public boolean getIsCheck() {
return isCheck;
}
public String getName() {
return name;
}
}
rule "Unique data"
no-loop
salience 10
when
$data1: Data( $id1: id,isCheck == false)
$data2: Data( isCheck == true, id == $id1)
then
log.log(" data are not unique:"+ $data1.getName() +" and
"+$data2.getName()+ " id is "+$data2.getId()+"
"+$data1.getId());
end
rule "Change data"
no-loop
when
$data: Data(isCheck == false)
then
$data.setIsCheck();
update($data);
end
So you can see from above that I set value false at the start and then set
it to true in a different rule. Now it takes only less then a second to
execute the rule. I was not aware of the fact that each time i insert a rule
it start executing it.
Now my question is
Is it possible to stop working memory to start processing when I am
inserting data ? So don't execute anything untill I am finished adding all
the data and then fireAllrules
thanks ( also thanks to Scott Reed-3 who led to this solution)
Gaurav
--
View this message in context:
http://www.nabble.com/simple-rule-takes-long-time-tp18192098p18233804.html
Sent from the drools - user mailing list archive at Nabble.com.
5:11 a.m.
Hi,
Rules are not executed as Facts are inserted into Working Memory.
Basically, think of the Working Memory as a network of connected classes
("The RETE network"). As facts are inserted into Working Memory they
enter the top of the network and propagate through it depending upon the
patterns within the left-hand-side of the rule (WHEN). Should all
pattern conditions be met an activation is queued on the agenda. When
fireAllRules is called the activations (right-hand-side; THEN) are
executed which might in turn lead to changes in the network and other
activations being created or existing removed.
So, as you insert Data objects the LHS are matched within the network
but the RHS's do not run until you call fireAllRules.
Have you tested your rules below and confirmed they provide the correct
results? I am not convinced.
With kind regards,
Mike
-----Original Message-----
From: rules-users-bounces(a)lists.jboss.org
[mailto:rules-users-bounces@lists.jboss.org] On Behalf Of ygaurav
Sent: 02 July 2008 10:48
To: rules-users(a)lists.jboss.org
Subject: Re: [rules-users] simple rule takes long time
Hi All
I have few changes to my data class as suggested above and rule and it
seems
to work. But that brings me to another question. Here is my new data
class
Data class
public class Data {
private int id =0;
public boolean isCheck = false;
public String name = "";
public Data(int id) {
this.id = id;
name = "name"+(int)(this.hashCode());
System.out.println("id is "+id+" name is "+name);
}
public int getId() {
return id;
}
public void setIsCheck() {
isCheck = true;
}
public boolean getIsCheck() {
return isCheck;
}
public String getName() {
return name;
}
}
rule "Unique data"
no-loop
salience 10
when
$data1: Data( $id1: id,isCheck == false)
$data2: Data( isCheck == true, id == $id1)
then
log.log(" data are not unique:"+ $data1.getName() +" and
"+$data2.getName()+ " id is "+$data2.getId()+"
"+$data1.getId());
end
rule "Change data"
no-loop
when
$data: Data(isCheck == false)
then
$data.setIsCheck();
update($data);
end
So you can see from above that I set value false at the start and then
set
it to true in a different rule. Now it takes only less then a second to
execute the rule. I was not aware of the fact that each time i insert a
rule
it start executing it.
Now my question is
Is it possible to stop working memory to start processing when I am
inserting data ? So don't execute anything untill I am finished adding
all
the data and then fireAllrules
thanks ( also thanks to Scott Reed-3 who led to this solution)
Gaurav
--
View this message in context:
http://www.nabble.com/simple-rule-takes-long-time-tp18192098p18233804.ht
ml
Sent from the drools - user mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
7:35 a.m.
The LHS of "Change data" is true for every initial datum so I believe
what is happening is that it sets all your isCheck flags true
immediately before the "Unique Data" rule executes. I probably wasn't
very clear about what the flag was supposed to do. I meant to suggest
using a flag to record that a duplicate was detected to prevent running
the rule on any data that is found to be a duplicate. "isDuplicate"
would probably be a better name for the flag.
All you should need is one rule like the following:
rule "Unique data"
no-loop
salience 10
when
$data1: Data( $id1: id )
$data2: Data( isCheck == false, id == $id1)
then
log.log(" data have same id (" + $data2.getId() + "):"+
$data1.getName() +" and"+
$data2.getName());
$data1.setIsCheck(); // $data1 is duplicate
update($data1);
$data.setIsCheck(); // $data is duplicate
update($data);
end
ygaurav wrote:
Hi All
I have few changes to my data class as suggested above and rule and it seems
to work. But that brings me to another question. Here is my new data class
Data class
public class Data {
private int id =0;
public boolean isCheck = false;
public String name = "";
public Data(int id) {
this.id = id;
name = "name"+(int)(this.hashCode());
System.out.println("id is "+id+" name is "+name);
}
public int getId() {
return id;
}
public void setIsCheck() {
isCheck = true;
}
public boolean getIsCheck() {
return isCheck;
}
public String getName() {
return name;
}
}
rule "Unique data"
no-loop
salience 10
when
$data1: Data( $id1: id,isCheck == false)
$data2: Data( isCheck == true, id == $id1)
then
log.log(" data are not unique:"+ $data1.getName() +" and
"+$data2.getName()+ " id is "+$data2.getId()+"
"+$data1.getId());
end
rule "Change data"
no-loop
when
$data: Data(isCheck == false)
then
$data.setIsCheck();
update($data);
end
So you can see from above that I set value false at the start and then set
it to true in a different rule. Now it takes only less then a second to
execute the rule. I was not aware of the fact that each time i insert a rule
it start executing it.
Now my question is
Is it possible to stop working memory to start processing when I am
inserting data ? So don't execute anything untill I am finished adding all
the data and then fireAllrules
thanks ( also thanks to Scott Reed-3 who led to this solution)
Gaurav
5991
days inactive
5993
days old
5 comments
3 participants
participants (3)
-
Anstis, Michael (M.) -
Scott Reed -
ygaurav