Several types of dimensions in your Business warehouse:
Conformed: dimension shared among several fact tables (e.g.: a single product categorization that is used by all business aspects).
Non-Conformed: dimension targeted to a single fact table dimension where part of attributes is used in some part of business and different attributes used in different part of business. Cross-overs are possible. (e.g.: a buying product dimension and sales product dimension, where different aspects of business are using different attributes of a same entity – in this case same product).
Shared: dimension is used by several fact tables. Dimension key is stored in fact table but the value is stored in dimension table with other attributes. It is used by multiple measures.
Degenerated: dimension is when dimension attribute is stored directly into fact table and not in dimension table. It has no specific key and no other attribute.
Slowly changing: dimension where attribute value is changed over time and the change is retained in the dimension table based on type of slowly changing dimension (CSD):
Type I : Attributes history is not retained but updated. E.g.: Customer moved from place A to place B and all purchases will be connected to place B; also purchases from place A. This is suitable for time series and forecasts.
Type II : Change in attribute value creates a new record; Two records are showing purchases connected to place A and purchases connected to place B. In data analysis, user has option of selecting based on type of analysis (historical vs. forecasting).
Type III : Change in attribute creates only one record; original and last (current or the latest) value to this attribute. E.g.: this record would hold the date and name of the first place; place A and the latest move to place B. If more moves have occurred between place A and B, only the latest would be stored, namely place B.
Time dimension, hierarchies, …