Ok, I think my initial explanation was not so gut. I will try to clarify the scenario.

Supose that in the Data Hub I have the following table with information about customer purchases:

Customer IdArticleDeparmentAmountDate of Purchase

545 Shoes Fashion 30$ 01.07.2012

545 Book Library 20$ 01.07.2012

545 CD Library 20$ 01.07.2012

From this table I load my first table with Aggregated data called Department Aggregations

Customer IdDepartmentAmount Date of Purchase

545 Fashion 30$ 01.07.2012

545 Library 40$ 01.07.2012

I query the table above and fill a second table with aggregated data

Customer IdTotal ArticlesTotal DeparmetsTotal AmountAverage Shopping Days

545 3 2$ 70$ 1

Finaly I load the Customer table:

Customer IdTelephoneEmail CityTotal expenses

545 654654 cust@exp.com ABC 70$

My question is: When our dear customer, Mr 545, makes a new purchase, what is the less expensive option in order to update all the tables? I think that an incremental approach would be good because the actual Fact_Details tables contains around 165 millions of records and truncate this table and loaded again is too expensives, but I don't know how to updates aggregations.

Any comment will be appreciated

Kind Regards