How to document in VP ER model to model transformation in most efficient way?

Part of data governance is not only modeling entities, but in data ware house you have multiple layers, which do kind of data transformation/normalization.

Is there any best way how to document these kind of processes on attribute entity level?

Usual task is:

  • apply function on data attribute
  • map single field from source model to target model
  • map multiple fields from source to target model and vice versa
  • lookup to external data source
  • etc.

So its about design of ETL task

I am quite stuck in this area about how to utilize VP in best way to document this model lineage transition.

Thank you for any response.

I am now evaluating VP for this task before obtaining license.

So I found answer myself, I think this paper demonstrates in quite robust way how to achieve ETL transformation design in pure UML way and I like that approach.

Sharing link for others interested:
https://www.researchgate.net/profile/Sergio_Lujan-Mora/publication/221269120_Data_Mapping_Diagrams_for_Data_Warehouse_Design_with_UML/links/0912f50647a5ae23a3000000/Data-Mapping-Diagrams-for-Data-Warehouse-Design-with-UML.pdf?origin=publication_detail

1 Like

Actually it is not an solution at moment, because to make it working I need to transit class to ER and vice versa. The current problem is that when I transit Class to ER or ER to Class, the attributes/columns are not inherited…

Is there any way how to do this?

So I am trying to use UML based mapping. The first step is to represent in Class diagram model to model mapping where package stereotyped as <> is a package which has linked subdiagram:

So, this works in VP, but when I click on package and enter the low-level mapping object/diagram, I need to represent Attribute level mapping, so each attribute must be represented as Class with defined StereoType, so it looks like this:

So you see there is every higher level Class attribute represented as Class with stereotype <>, so we can continue on mapping low granularity.

Question: Can we extract Attributes from Class in form of another Class so we can:

  • prevent manual definition
  • automatic synchronization with class model
    ?

Or How difficult it would be to drag and drop attribute from Class in Class diagram and generate new Element with <> stereotype and with name Class.attribute… I think I can write some plugin for this…

I found:

But I don’t know how to work with drag and drop actions in the diagram to achieve my goal, any kind of code snippets would be useful.

Thank you for any hints. With this supported VP can become truly Data Mapping Tool.

Best regards,

Ladislav

I welcome also other approaches to this topic… other than what I introduce…

How about use Composite Structure Diagram to show the attributes as Part/Property shapes,
and shows the mapping between the Part/Property shapes?

PS: The model elements of Class & Attributes are shared in Class Diagram & Composite Structure Diagram. It is unnecessary to ‘synchronize’ the class models.

1 Like

@peter.wong - this is not bad, really not bad :slight_smile:

Thanks a lot, going to investigate by hand myself!

I have entities with over hundred of attributes, it will be handy if I can simply select ALL attributes to show on the class, otherwise its very long management task.

1 Like

Exactly, also I should be able to make the attributes being aligned vertically/horizontally within Class.

Your enhancement request “show all attributes” is passed to our engineer team.
We will notify you once it is supported.

Unfortunately, align attributes vertically / horizontally within Class is not yet supported.

I would like to let you know our engineers had deployed a new build which supported “show all attributes”.

Please update the software to latest patch build to get this fix.
Details about update to latest patch can be found at

1 Like

Thx a lot guys! This will simplify a lot my POC for using VP as base for documenting the data integration flows. Ultimately I would like to use VP as source for generating the Huawei Cloud data integration pipeline.