SAP Star Schema
1. The Classic Star Schema1) 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
页:
[1]