Eclipse Community Forums: Epsilon

Home

Home » Modeling » Epsilon » ETL

ETL [message #510015]

Tue, 26 January 2010 03:39

Eclipse User

Hello again. I have the following problem: when there is executed the following rule of transformation of primitive types the problem that I have is that when a type does not expire with the rule, in this case that is of type "DATED" does not expire with the rule String2Varchar there is generated a primitive type that for fault is of type CHAR.

rule String2Varchar
transform pr:UML!PrimitiveType
to v:modeloOR!PrimitiveType{

if((pr.name='String') or(pr.name='STRING') or (pr.name='string')){
'String2Varchar'.println();
v.Name:=modeloOR!BasicTypes#CHAR;
}
}

But I believe that the correct functioning is that if a type for example to type "DATE" it does not expire with the "if", it does not have to generate a TypePrimitive on not having expired with the judgment IF.

Dimitros, your tell me that was creating a file '.model ' but ¿ I can it execute the rules ETL from a file '.ecore ' of entry and another file '.ecore ' of exit?

If the file of entry was '.ecore ', the rules ETL serian like that?

rule Package2Model
transform t : ecore!Package
to n : ecore!Model {
n.Name:= t.Name;
'Package2Model '.println();
}
Thanks!!!

Re: ETL [message #510038 is a reply to message #510015]

Tue, 26 January 2010 00:20

Eclipse User

Hi,

Please see comments below.

c wrote:
> Hello again. I have the following problem: when there is executed the
> following rule of transformation of primitive types the problem that I
> have is that when a type does not expire with the rule, in this case
> that is of type "DATED" does not expire with the rule String2Varchar
> there is generated a primitive type that for fault is of type CHAR.
>
>
> rule String2Varchar
> transform pr:UML!PrimitiveType
> to v:modeloOR!PrimitiveType{
> if((pr.name='String') or(pr.name='STRING') or (pr.name='string')){
> 'String2Varchar'.println();
> v.Name:=modeloOR!BasicTypes#CHAR;
> }
> }
>
>
>
>
> But I believe that the correct functioning is that if a type for example
> to type "DATE" it does not expire with the "if", it does not have to
> generate a TypePrimitive on not having expired with the judgment IF.
>
>

I'm afraid I can't understand your question. Could you possibly rephrase
and/or attach some model files with the expected/actual results of the
transformation?

>
> Dimitros, your tell me that was creating a file '.model ' but ¿ I can it
> execute the rules ETL from a file '.ecore ' of entry and another file
> '.ecore ' of exit?
>
> If the file of entry was '.ecore ', the rules ETL serian like that?
>
>
> rule Package2Model
> transform t : ecore!Package
> to n : ecore!Model {
> n.Name:= t.Name;
> 'Package2Model '.println();
> }

If you want to transform one Ecore model to another Ecore model, you
have to give distinct names to your models (e.g. Source and Target) and
make sure you use Ecore classes in your transformation (Ecore doesn't
define a Package or Model class). E.g.

rule EPackage2EPackage
transform s : Source!EPackage
to t : Target!EPackage {

t.name = s.name;
...

}

> Thanks!!!

Cheers,
Dimitris

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

Re: ETL [message #510069 is a reply to message #510038]

Tue, 26 January 2010 08:10

Eclipse User

I have the following function:

rule String2Varchar
transform pr:UML!PrimitiveType
to v:modeloOR!PrimitiveType{
if((pr.name='String') or(pr.name='STRING') or (pr.name='string')){
'String2Varchar'.println();
v.Name:=modeloOR!BasicTypes#CHAR;
}
}

And in the model of entry I have the types

Primitive integer
Primitive date
Primitive STRING

The resultant model is
Primitive CHAR
Primitive CHAR
Primitive CHAR

It is to say that the type Primitive integer turns into Primitive CHAR

I do not understand because it happens
Thanks!!!

Re: ETL [message #510203 is a reply to message #510069]

Tue, 26 January 2010 11:35

Eclipse User

Hi,

This happens because CHAR is the first literal in the BasicTypes
enumeration and is used by EMF as the default value in case you don't
explicitly specify one.

Cheers,
Dimitris

c wrote:
> I have the following function:
> rule String2Varchar
> transform pr:UML!PrimitiveType
> to v:modeloOR!PrimitiveType{ if((pr.name='String') or(pr.name='STRING')
> or (pr.name='string')){
> 'String2Varchar'.println();
> v.Name:=modeloOR!BasicTypes#CHAR;
> }
> }
>
>
> And in the model of entry I have the types
>
> Primitive integer
> Primitive date
> Primitive STRING
>
> The resultant model is
> Primitive CHAR
> Primitive CHAR
> Primitive CHAR
> It is to say that the type Primitive integer turns into Primitive CHAR
>
> I do not understand because it happens
> Thanks!!!

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

Re: ETL [message #511239 is a reply to message #510015]

Sun, 31 January 2010 11:36

Eclipse User

I have some doubts:
The first one of the doubts that tense is:
I have the code:
operation method_seq(c:UML!Class):modeloOR!Method{
var m: new modeloOR!Method;
for (op in c.ownedOperation) {
m.Name:=op.name;
}
return m;
}

operation asignartt(c:UML!Class):modeloOR!TypedTable{
'asignartt'.println();
var sx: new modeloOR!TypedTable;
sx.Name:=c.name+'s<<persistent>>';
return sx;
}

rule Class2UDT
transform c:UML!Class
to s:modeloOR!StructuredType{

'Class2UDT'.println();
s.Name:=c.name+ "<<udt>>";
s.Name.println();

s.typed:=asignartt(c);
s.method:= method_seq(c);
}

And for the assignment of s.typed: = asignartt (c); and s.method: = method_seq (c); me the mistake appears " Invalid assignment to property 'typed' " and " Invalid assignment to property 'method' " and I do not understand because this happens.

The second doubt:
When I transform a type CHAR the method creates
Primitive CHAR
Primitive CHAR
Primitive CHAR
and I want that the result is
Primitive integer
Primitive date
Primitive STRING
You were recommending to your me that ' you don't explicitly specify one.
But I do not want to create anything type for default. I want that alone when the condificion 'if' is true it create the type PrimitiveType CHAR.
To obtain it I have thought of putting a precondicion in front of this method of the following form:
pre{
var pr: new UML!PrimitiveType;
(pr.name='String') or(pr.name='STRING') or (pr.name='string')
}
rule String2Varchar
transform pr:UML!PrimitiveType
............
...........

But this does not work. Since I can do it?

The third doubt is, which is the equivalence of the method resolveTemp" of ATL in ETL.
The defincion of this function in ATL is:

This specific operation makes it possible to point, from an ATL rule, to any of the target model elements (including non-default ones) that will be generated from a given source model element by an ATL matched rule.
The operation resolveTemp has the following declaration: resolveTemp(var, target_pattern_name). The parameter var corresponds to an ATL variable that contains the source model element from which the searched target model element is produced. The parameter target_pattern_name is a string value that encodes the name of the target pattern element (see Section 4.5.2) that maps the provided source model element (contained by var) into the searched target model element.

I have this code:
typed <-thisModule.resolveTemp(p->refImmediateComposite(),'tt')

I have thought that the transformation would be in ETL:

operation resolveTemp(d:UML!Property):modeloOR!TypedTable{
'resolveTemp'.println();
var t: new modeloOR!TypedTable;
d.eContainer():='tt';
return t;
}

rule.xxxx
transform :d:UML!Property
to .....
{
st.typed:=resolveTemp(d);
}
but, it is not correct....And I also have tried this:
typed:=(p.eContainer().type::='tt');

Thanks!!

Re: ETL [message #511256 is a reply to message #511239]

Sun, 31 January 2010 17:53

Eclipse User

Hi,

Could you please send me a complete snapshot of your project
(transformation, models, metamodels etc) so that I can have a look at this?

Cheers,
Dimitris

c wrote:
> I have some doubts:
> The first one of the doubts that tense is:
> I have the code:
> operation method_seq(c:UML!Class):modeloOR!Method{
> var m: new modeloOR!Method; for (op in c.ownedOperation) {
> m.Name:=op.name;
> }
> return m;
> }
> operation asignartt(c:UML!Class):modeloOR!TypedTable{
> 'asignartt'.println();
> var sx: new modeloOR!TypedTable;
> sx.Name:=c.name+'s<<persistent>>';
> return sx;
> }
>
> rule Class2UDT
> transform c:UML!Class
> to s:modeloOR!StructuredType{
> 'Class2UDT'.println();
> s.Name:=c.name+ "<<udt>>";
> s.Name.println();
> s.typed:=asignartt(c);
> s.method:= method_seq(c); }
>
> And for the assignment of s.typed: = asignartt (c); and s.method: =
> method_seq (c); me the mistake appears " Invalid assignment to property
> 'typed' " and " Invalid assignment to property 'method' " and I do not
> understand because this happens.
>
>
> The second doubt:
> When I transform a type CHAR the method creates Primitive CHAR
> Primitive CHAR
> Primitive CHAR and I want that the result is
> Primitive integer
> Primitive date
> Primitive STRING
> You were recommending to your me that ' you don't explicitly specify one.
> But I do not want to create anything type for default. I want that
> alone when the condificion 'if' is true it create the type PrimitiveType
> CHAR.
> To obtain it I have thought of putting a precondicion in front of this
> method of the following form:
> pre{
> var pr: new UML!PrimitiveType;
> (pr.name='String') or(pr.name='STRING') or (pr.name='string')
> }
> rule String2Varchar
> transform pr:UML!PrimitiveType
> ...........
> ..........
>
> But this does not work. Since I can do it?
>
>
>
> The third doubt is, which is the equivalence of the method resolveTemp"
> of ATL in ETL.
> The defincion of this function in ATL is:
>
> This specific operation makes it possible to point, from an ATL rule, to
> any of the target model elements (including non-default ones) that will
> be generated from a given source model element by an ATL matched rule.
> The operation resolveTemp has the following declaration:
> resolveTemp(var, target_pattern_name). The parameter var corresponds to
> an ATL variable that contains the source model element from which the
> searched target model element is produced. The parameter
> target_pattern_name is a string value that encodes the name of the
> target pattern element (see Section 4.5.2) that maps the provided source
> model element (contained by var) into the searched target model element.
>
>
> I have this code:
> typed <-thisModule.resolveTemp(p->refImmediateComposite(),'tt')
>
> I have thought that the transformation would be in ETL:
>
> operation resolveTemp(d:UML!Property):modeloOR!TypedTable{
> 'resolveTemp'.println();
> var t: new modeloOR!TypedTable;
> d.eContainer():='tt';
> return t;
> }
>
> rule.xxxx
> transform :d:UML!Property
> to .....
> {
> st.typed:=resolveTemp(d);
> }
> but, it is not correct....And I also have tried this:
> typed:=(p.eContainer().type::='tt');
>
> Thanks!!
>
>
>
>

Re: ETL [message #511871 is a reply to message #511256]

Wed, 03 February 2010 05:05

Eclipse User

Can someone help me?
Thanks..

Re: ETL [message #512049 is a reply to message #511871]

Wed, 03 February 2010 12:57

Eclipse User

I've replied to your email.

Cheers,
Dimitris

c wrote:
> Can someone help me?
> Thanks..

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

Re: ETL [message #512112 is a reply to message #512049]

Wed, 03 February 2010 18:07

Eclipse User

Thank you for answering me. I am sorry to be a nuisance so much, but I try to solve the problems for hours before writing in the forum.
I know that the error that appears of 'Invalid assignment to property 'typed' " ' refers to incompatible types,but this is not true.
Before writing in the forum, use the functions '.println () ' of the following form
//Invalid assignment to property 'typed'
//s.typed:=asignartt(c);
'type is'.println();
asignartt(c).println();'

and the exits is:

'type is
TypedTable [Name=Class2UDTs<<persistent>>, ] '.

But I look that the type of the variable 's.typed' is of the same type that the value returned by the function'operation asignartt(c:UML!Class):modeloOR!TypedTable{'.
In the file' '
I look in the file 'modeloOR.ecore' at the field 'typed' of the StructuredType that is the type of the variable 's' and also it is of type TypedTable

Because of it I have written in the forum. Because it gives a problem of incompatible types when it is not true and do not be to which other reasons it can owe.
Thank you for saying the existence of the function to me 'equivalent' and excuse me to my for the inconveniences
A greeting

Re: ETL [message #512320 is a reply to message #512112]

Thu, 04 February 2010 11:37

Eclipse User

Hi Carol,

I can't reproduce this with the last version of the transformation you
sent me (1st Feb). After fixing the name of UM.model -> UML.model,
launch.xml runs with no errors... Perhaps you've sent me an older version?

Cheers,
Dimitris

c wrote:
> Thank you for answering me. I am sorry to be a nuisance so much, but I
> try to solve the problems for hours before writing in the forum.
> I know that the error that appears of 'Invalid assignment to property
> 'typed' " ' refers to incompatible types,but this is not true.
> Before writing in the forum, use the functions '.println () ' of the
> following form //Invalid assignment to property 'typed'
> //s.typed:=asignartt(c);
> 'type is'.println();
> asignartt(c).println();'
>
> and the exits is:
>
> 'type is
> TypedTable [Name=Class2UDTs<<persistent>>, ] '.
>
> But I look that the type of the variable 's.typed' is of the same type
> that the value returned by the function'operation
> asignartt(c:UML!Class):modeloOR!TypedTable{'.
> In the file' '
> I look in the file 'modeloOR.ecore' at the field 'typed' of the
> StructuredType that is the type of the variable 's' and also it is of
> type TypedTable
>
> Because of it I have written in the forum. Because it gives a problem of
> incompatible types when it is not true and do not be to which other
> reasons it can owe.
> Thank you for saying the existence of the function to me 'equivalent'
> and excuse me to my for the inconveniences
> A greeting

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

Re: ETL [message #512406 is a reply to message #512320]

Thu, 04 February 2010 18:25

Eclipse User

Hello Dimitrios. I have sent the last version to your mail.The reason of that your you do not see the mistakes is because the line this of code commented.

Thank you.

Re: ETL [message #512742 is a reply to message #512406]

Sat, 06 February 2010 20:00

Eclipse User

Hi Carol,

OK. I've now reproduced this. The problem is that StructuredType.typed
is an 1..* feature but asignartt() returns a single TypedTable.
Therefore, instead of performing an assignment

s.typed:=asignartt(c);

you should be adding the result to s.typed:

s.typed.add(asignartt(c));

Fixing this makes the transformation run with no errors.

Cheers,
Dimitris

c wrote:
> Hello Dimitrios. I have sent the last version to your mail.The reason of
> that your you do not see the mistakes is because the line this of code
> commented.
>
> Thank you.

Re: ETL [message #585395 is a reply to message #511239]

Sun, 31 January 2010 17:53

Eclipse User

Re: ETL [message #585617 is a reply to message #511256]

Wed, 03 February 2010 05:05

Eclipse User

Can someone help me?
Thanks..

Re: ETL [message #585649 is a reply to message #585617]

Wed, 03 February 2010 12:57

Eclipse User

Re: ETL [message #585690 is a reply to message #512049]

Wed, 03 February 2010 18:07

Eclipse User

Re: ETL [message #585703 is a reply to message #585690]

Thu, 04 February 2010 11:37

Eclipse User

Re: ETL [message #585715 is a reply to message #512320]

Thu, 04 February 2010 18:25

Eclipse User

Hello Dimitrios. I have sent the last version to your mail.The reason of that your you do not see the mistakes is because the line this of code commented.

Thank you.

Re: ETL [message #585815 is a reply to message #585715]

Sat, 06 February 2010 20:00

Eclipse User

Previous Topic:	[EOL, Ecore2GMF.eol] How duplicate GmfMap!NodeMapping?
Next Topic:	Transform UML Class diagram to UML Class diagram

Goto Forum:

-=] Back to Top [=-

Current Time: Tue Jul 22 19:18:55 EDT 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter