Introduce 1 to many table synchronisation in Design Layers


1 comment

  • Avatar
    Jason Shangold
    • I would like the whole design layer approach to be looked at for revisions. I would especially like the ability to be able to ask of a source model,  who are all my children (derived child models).  Since the child holds the information as to its parentage, it is not possible to report directly on which models wee derived from  particular model.

      • TheovanWestrienen

        Yes, I agree, I would like a lot more too. But: connecting between models is technically possible (even thought quite cumbersome), whereas this 'idea' is really adding unexisting functionality to ERwin.

        I see 2 ways of reporting on source to target (i.e. derived information)

        - building queries using ODBC driver ( which I am currently experimenting with this; doable but serious amount of work involved)

        - deploying the ERwin webportal: built in where used functionality, including a graphical presentation of model lineage.


        So I did choose to limit my idea to the functionlity I need, and for which no automated workaround is available at this moment


        • Edwy_Erwin

          if you don't have a lot of tables falling into this scenario, you better off replicate the table that has columns from different tables in your logical and use it to sync.

          actually I never tried but can you create a view in the logical that has the columns from whatever table you want and make it the source of your table on the physical models.


      I totaly agree to that idea. This is very big gap of functionality in ERwin!

      I am working 100% in data warehouse environments. Therfore, I have a conceptional business model which is presented within ERwin as a logical model. The physical model is design as a data vault model. This means that a entity in a logical model may decomposed into three different data-vault-entity types. Currently, there is no way to synchronize the logical ERwin model to the physical data vault model.


      Some notes regarding the data lineage (that's a different topic in my opinion): In all my projects I reverse engineer the source-systems model and save it as a ERwin model. Then I enriche the model with metadata taken as a result of the busness analysis. Then I model the next layer of my data warehouse arcitecture, e.g. staging area. I import the source-system-model as a new DataWarehouseSource into the Staging-Area-Model to create a logical source-to-target-mapping. I repeat these steps for each BI-layer of my data warehouse. Last step is the graphical representation of the data-lineage via the web-portal.

    Comment actions Permalink

Please sign in to leave a comment.

Powered by Zendesk