http://www.kimballgroup.com/2012/02/design-tip-142-building-bridges/
The dominant topologies of Data Warehouse modelling (Star, Snowflake) are designed with one-to-many relationships in mind. Query readability, performance, and structure degrades severely when faced with a many-to-many relationship in these modelling schemes.
? ?
1. Type2 Fact tables, how to let the BI report tools like BO, micro strategy support automatically extract correct date range data?
? ?
2.if one row in fact table has more than one row in the dimension tables, what should we do?
Normalize the fact table is a bad idea. It will cause your measures use more distinct for getting correct result.
Use a bridge table to join the fact table and dimension table.
? ?
3.add the bitmap index for the stage tables, if these tables we will take more batches. It will be more effective than B plus tree index in this case.
http://docs.oracle.com/cd/E11882_01/server.112/e25554/indexes.htm#DWHSG8131
Design Tip #142 Building Bridges
原文：http://www.cnblogs.com/huaxiaoyao/p/4242085.html