Hi eb
I have come accross 3 issues concerning ROLAP dimensions while reading forums discussions and would appreciate your thoughts about them:
1)ROLAP dimensions processing - Assuming full data integrity in cube's star schema I understand there is no point in doing additional dimension processing since there is no need to handle attributes key errors or maintain some intrinsic structure for such a dimension in SSAS at all(??)
2)Query perforemance - should it slow down significantly at list when a ROLAP dimension cube's cache is cold (because dimension's structure needs to be built on the fly from underlying dimension tables ??)
3)data integrity - assuming it is not perfect on underlying star schema tables and no processing of ROLAP dimensions was done could i still use dimension's and Partition's error configuration to handle data integrity issues or it would be useless in this case?
For example use SSAS unknownMember as solution for errors...??
4)If processing of ROLAP objects is the only opportunity to handle attribute keys errors - what happens then with FATCS keys errors (missing keys,null keys etc) of ROLAP partitions which are never processed??
TIA
Rea