SAP Star Schema

sqlo

　　1. The Classic Star Schema
1) Dimension Tables
    Dimension attribute/Described attributes(non-dimension attributes)
    answer to "what/whe/who/..."
    Described attributes correspond to master data.      
2) Fact Table
    answer to "how much/how many", focus on "amount/quantity..."
    link to dimension table by a unique key
    From OLTP perspective, fact table <--->transaction data      
3) Star Schema can answer these questions:
    who have sold to?
    what have we sold?
    how much have we sold?
    when did we sell it?
   
2. The SAP Star Schema
1) Problems of the classic star schema:
   --不能灵活的创建Drill Down分析
   --支持多语言很困难
   --在大表上(Fact Table)建立基于字符字段的索引，导致数据访问性能较差
   --如果attributes of dimensions随着时间变化，不容易维护旧值和新值的对比分析
   --Master data被多次复制，需要较多的Infocube
   --数据的层次关系(Hierarchy relationship)需要通过建模成attributes of dimension来实现  
2) SAP BW Star Schema
   --Key Figures-->Facts of Fact Table
   --Characteristics-->Dimension attribute
   --Dimension Table(Optional)/SID Tables/(Optional)Master Data Tables-->Dimension(Dimension Table)
       但是characteristics不存在Dimension table中，而是使用一个替代ID(SID)来对应于characteristics
       存放于Dimension table中
   --Additional information about characteristics被存放在Master Data Table中，
       Attributes
       Texts
       External Hierarchy