Aggregation for Data Warehouse, Part 5
Posted by Dylan Wan on June 4, 2007
This is the final posting about the aggregation for data warehouse. In the prior posts, I described the following approaches of providing aggregation for data warehouse.
I think that for a large scale data warehouse project, a two tiered approach may be favorable. When I say a two tiered approach, I mean that you can have a data warehouse, and …
You can have a data warehouse and have OLAP cubes as your data marts. This is actually the approach I learned from a Hyperion white paper – Large-Scale Data Warehousing Using Hyperion Essbase OLAP Technology.
The problem of building aggregation tables is that you need to know which level the data is typically reported on so you can create specific aggregation at those levels. Using a OLAP Cube, you can avoid this problem since the summarized data are available at all different levels.