Hi,
The 2 approaches are as follows:
Aggregate tables are built in the database, which contains the dimension fields(not foreign keys) along with the aggregated measures. In the universe they are present as standalone tables, i.e they are not joined with any dimensions. Aggregate aware function is used to define both the dimensions and measures of such tables
No aggregate tables are built in the database level. They contain the normal fact table at different granularities. In the universe, aggregate aware is used only to define the measures and aggregate incompatibility is set accordingly
The first approach is better in terms of performance, since for the higher levels of aggregation, all the information is obtained for a single table. However, a large scale implementation of this approach in a dimensional schema is difficult. In most BI projects, the second approach is preferred
Regards,
Valluvan N
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.