Forum Stats

  • 3,838,691 Users
  • 2,262,394 Discussions


ODM: Double relationshipnames when using gitHub

We are using gitHub to version-control our model.

If 2 users in two clones of the same model create a relation, it happens, that the names of the relations are identical ("Relation_nnn", nnn incremented automatically). After commit, push and merge, we have two relations (in two files, different ids) with the same name. The modeler handles this by adding "v1" for the display to one of the names. BUT it does not store this unless you change something in the relation.

The same could happen with every element, having unique names.

This inconsistency (2 identical relation-names) remains in the xml-files. As we analyse these files for further usage, we run into trouble.

Any idea, how we can avoid duplicate names?


  • Philip Stoyanov-Oracle
    Philip Stoyanov-Oracle Member Posts: 3,420 Employee

    Any idea, how we can avoid duplicate names?

    1) better coordination in the team - split the model on subject areas and different people to work on different areas. As you pointed that could happen for any new object if there is another object with the same name and type(entity, table...) created in another branch.

    2) git functionality is acting first on merge. If there is a conflict then our code is acting (local merge) in order to minimize damages. If there are no conflicts (according git) then DM cannot do anything. You can use two local clones of repository for different branches and use DM functionality for 'compare and merge models" or import of DM design. Then you can analyze reported changes and decide whether to go with Git merge or use DM import