In the original: "Bi Thing" Data Warehouse Modeling: Star mode, snowflake pattern
Data Warehouse Modeling-Star schema
Example of Star Schema
Data warehousing Modeling-snowflake mode
Example of Snowflake Schema
- Save storage space
- A certain degree of paradigm
Star vs. Snowflake type
- A long debate
- There are supporters of both views
- The argument continues ...
- For now, most of them are more likely to be star-shaped
Arguments that support star dimensions
- The fact table is always big, and the space saved on the dimension table is relatively small.
- Increased complexity of the data model
- Query operations are conceptually more complex
- Load time from Data Warehouse to multidimensional database is longer
- Therefore, the Snowflake dimension is considered only when the dimension table is extremely large and the storage space is a problem
- In short, it's best to use the Star dimension
Arguments to support snowflake dimensions
- From the data warehouse to the multi-dimensional database loading process, the snowflake dimension is more efficient;
- The Snowflake dimension describes a clearer hierarchical concept;
- Stars are considered only when the end user is likely to have direct access to the Data warehouse (which is not recommended);
My personal experience
- The star structure is better than the snowflake type in efficiency;
- In most cases, I will choose the star type, but do not rule out the use of snowflake-shaped case;
"Bi Thing" Data Warehouse Modeling: Star-mode, snowflake-mode