Attribute relationship play an important role on cube performance , it tell the SSAS engine how the aggregation on the cube dimension will be done.Recently i came across an issue where we have defined the attribute relation within one of our dimension as Rigid. At the time of defining this we didn’t gave a thought if the dimension data will be updated or deleted.
Later on for the sake of improving Cube performance we changed the Processing type for all the dimension to Update Process , and there we hit an error saying “Rigid relationships between attributes cannot be changed during incremental processing of a dimension“.
This happens because the data has changed within our dimension and we defined our Attribute relationship as Regid, which by right should be defined as Flexible, another reason for this error was the Dimension processing type which was Update Process. In order to resolve this issue you may either change the attribute relationship to Flexible or change the Dimension Processing type to Full Process.
The point to note here is that when you define your Dimension Processing type as Update Process and the dimension attribute relationship is defined as Rigid, SSAS expects that the dimension data will remain static and the attribute relationship will not change.