Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Guillaume

Pages: 1 ... 4 5 [6] 7 8 ... 37
General Board / Re: Associations order in the generated XSD file
« on: December 09, 2016, 06:50:14 pm »
Hi Simon,

Sounds like what I'm looking for!
Can you advise the tagged value that needs to be created on the associations?
I tried to add position and order = 1, 2, 3... but it didn't work.


Bugs and Issues / Re: SysML 1.3 IBD and EA13
« on: December 09, 2016, 06:42:13 pm »
Hi Simon,

Thanks for the clarification.

General Board / Associations order in the generated XSD file
« on: December 09, 2016, 03:11:55 am »

I have a project where we defined XSDcomplexType classes to generate XSD files.
Using EA Generate XML Schema command, here is an extract of what I get on the associations created between complex types:
Code: [Select]
<xs:complexType name="ComplexType1">
                    <xs:element name="element1" type="xs:string" minOccurs="1" maxOccurs="1"/>
                    <xs:element name="association1" type="Association1" minOccurs="1" maxOccurs="1"/>
                    <xs:element name="association2" type="Association2" minOccurs="1" maxOccurs="1"/>
                    <xs:element name="association3" type="Association3" minOccurs="1" maxOccurs="1"/>
                    <xs:element name="association4" type="Association4" minOccurs="1" maxOccurs="1"/>
Having looked into details, EA uses the created order i.e. Connector ID value from t_connector table (I get association 4 before association 3, if I created associations in the order 1 / 2 / 4 / 3).
So the I'm looking for a way to let EA users change this order.

Could the Generate XML Schema in Sparx EA be improved to support that e.g. via tagged value or else?
I tried the Schema Composer which seems to apply an alphabetical order on the role name only. I also noticed that the Schema Composer doesn't generate an XSD content as complete as EA Generate XSD file.
I'm aware that a Schema Composer addin can be written to generate an XSD file and support additional tagged values. I did something similar to generate JSON files before, but the current context requires an option that's built in EA.

Could it be looked at by Sparx?


Bugs and Issues / SysML 1.3 IBD and EA13
« on: December 08, 2016, 11:38:13 pm »

I noticed 2 weird behaviours in EA 13 (build 1309) and SysML IBD (SysML 1.3)
I created an IBD from a block; this diagram is automatically created under the block and I can see a frame view of the block.

Once I move this diagram to a separate package, this "frame" is replaced with the block itself ; how can I prevent this? (note that if I moved it back under the block, it's ok)

I also found that port names on an IBD can be blurry when you first open the diagram (it eventually goes back to normal).


General Board / Re: Link labels and text fields to class diagram attributes
« on: December 06, 2016, 08:57:05 pm »
Hi Matthias,

You need to use the Link to Element Feature - once you've created a connector, right click on the end (class side) and select Link to Element Feature. You can select the attribute to associate (see

I posted an article that contains such case here: (french version available here :

General Board / Re: PermissionID Description
« on: November 30, 2016, 12:35:31 am »
I recently shared the description via the article below:

QWERTY, I'm happy for you to include this list in the next update of your Inside EA ebook


Bugs and Issues / Re: stereotyped SysML blocks and IBD
« on: November 23, 2016, 06:32:08 pm »
Hi Simon,

In my UML profile, a stereotype extends the metaclass Class and has a generalization link to the SysML 1.3 Block stereotype.
This is my understanding of extending SysML, BPMN and other stereotypes. Is there a better way to achieve this?

Have you got a UML profile sample with a stereotyped SysML Block that works fine in IBD that you could share?

Bugs and Issues / stereotyped SysML blocks and IBD
« on: November 23, 2016, 08:11:57 am »

I defined a UML profile with stereotypes on SysML blocks.
Adding an IBD on those blocks work ok (sync element show properties and ports).

However when I add a stereotyped block as a "property" (SysML 1.3 Part), nothing happens.
If I create a normal SysML Block and then add / select my custom stereotype, then it works ok. It looks like the expected behaviour is obtained with the SysML 1.3 Block is selected as the main stereotype, and any custom stereotype must be added. This is not ideal as the associated toolbox doesn't achieve this.

Is there any this could be fixed? Any comment?


General Board / Re: ArcGIS Schema View export not working
« on: November 21, 2016, 08:24:57 pm »

Did you run the export with a right click on the "ArcGIS SchemaView" stereotyped package > ArcGIS > Export to ArcGIS Workspace.
I tested a while ago and it worked ok.

I believe this feature was covered in the following webinar:

General Board / Re: How to hide packages from users / user groups?
« on: November 16, 2016, 07:10:47 pm »
Some of my clients have requested a feature that hides packages in the Project Browser from users due to their EA model structure.
In one situation, the EA project had model roots for each "product": there was a clear need to hide model roots from user groups (there was reasons and constraints for running such structure)/
For another situation, there was a need to hide packages so stakeholders and reviewers could use EA to review the model branch that was relevant to them.

So I think there are purposes for such feature, however I didn't get a positive answer from Sparx on my request.

For the latter case (review), I think a tool with a web interface could easily let one manages what needs to be visible for certain users.

General Board / Re: Version Control While Accessing Model Via Citrix
« on: November 16, 2016, 07:01:37 pm »

Have you looked at installing the Cloud Service and enabling EA security to manage your user locks?
Sparx Cloud Service includes the WAN Optimizer which aims at improving performances with the DB (e.g. Oracle).
It also simplifies the clients' configuration as a url replaces the ODBC configuration.

Citrix Server is only useful if you don't want to install EA on each user's PC and need an optimised/central way of managing your licences (if you have floating ones).

I would suggest dropping Version Control as it quickly lead to slow access for users. The only advantage of Version Control compared with the EA Security is to provide an Undo Checkout (there are workaround to achieve this).
Version control is not great at maintaining history since controlled packages often change during the model life; hence data can move from a controlled XML file to another which can make it difficult use when needed.
There are options to keep history of your models: baselines, scheduled tasks to run an XMI export of your models, DB backup, audit logs...

I'm running EA 12.1 and since I moved to Windows 10, I'm getting issues with the standby mode.
When resuming, I'm getting a Busy Application popup with the action to retry ; I need to click several times before EA works again.

General Board / Re: Enterprise Architect v13 upgrade guide
« on: November 08, 2016, 07:05:08 pm »
Hi Simon,

I reported it as a bug.


General Board / Re: Enterprise Architect v13 upgrade guide
« on: November 07, 2016, 07:21:57 pm »
Quick questions from a client ref the DB Schema upgrade:

1- the Postgres schema upgrade creates 2 indexes on the public.t_files table - see:

CREATE INDEX ix_files_appliestonamesize ON public.t_files (appliesto ASC,name ASC,filesize ASC)
CREATE INDEX ix_files_appliestonamedatesize ON public.t_files (appliesto ASC,name ASC,filedate
ASC,filesize ASC)
Is this a typo? what is the reason for changing the public DB?

2-is the schema stored somewhere so one can check what version is currently applied on a running EA project? (without having to run a db schema compared)



I'm working on a custom MDA transformation template with simple rules from UML Class diagram to a DDL Data Model.
The attributes transformations to perform are:
- length = UML attribute's upper bound --> length=%qt%%attUpperBound%%qt% (works ok)
- column's PK = true if UML attribute's  isId = true
- column's Not Null = true if UML attribute's  lower bound > 1

I'm struggling in finding the appropriate fields to set the latter two.. Note that the PK is stored in "t_attribute.isordered" and Not Null is stored in  "t_attribute.alloduplicates"
Any help will be greatly appreciated.


Pages: 1 ... 4 5 [6] 7 8 ... 37