What is a Junk Dimension? Have you ever come across a scenario where many small dimension tables connect (using foreign key) to the fact table? In data warehouse, we have two types of tables, Fact and Dimension. Fact table contains business facts as measures and references to the dimension tables. Dimension tables have attributes which contains textual information of business facts and are used to filter and label the data.
Typically, fact tables are deeper whereas dimension tables are wider. Sometimes it happens that we have many narrow and low cardinality (low cardinality : few number of rows in the table) dimensions in the system; like flags and indicators. To keep all these information in fact table, we need to connect all these dimension tables (available in system or created from flag and indicator attributes) with fact table. Connecting all these small dimension tables converts fact table in … More