What is the star mode?
The star mode can be described as a simple star: the central table contains fact data, multiple tables are radial distribution centered on the central table, and they are connected to each other through the primary key and foreign key of the database.
What is a snowflake pattern?
The snowflake pattern represents a dimension model, which is also composed of a central fact table and a group of dimension tables. These dimension tables can be further normalized as subdimension tables.
When will the snowflake pattern be used for implementation?
Data Warehouse expert Ralph Kimball recommends three cases. In these three cases, not only can the snowflake be used for implementation, but it is also the key to successful design:
|
Large customer dimension, where, for example, 80% of fact table measurements involve anonymous visitors (You know little about their information), 20% involving customers registered in a reliable manner (You can track multiple dimensions to collect detailed data about them) |
|
Financial product dimensions of banks, brokers, and insurance companies, because each product has many special properties that cannot be shared with other products. |
|
Multi-enterprise calendar dimension, because each organization has a special checkout period, quarter, and holiday |
Ralph Kimball recommends that the star mode is a good solution in most cases. Although redundancy is reduced in the standard snowflake pattern, more connections are needed. Kimball generally recommends that you do not expose end users to a physical snowflake design because it always affects comprehensibility and performance.