Relational databases are made up of relations (related tables). Tables are made up of columns. If the tables are two large (i.e. too many columns in one table), then database anomalies can occur. If the tables are two small (i.e. database is made up of many smaller tables), it would be inefficient for querying. Normalization and Denormalization are two processes that are used to optimize the performance of the database. Normalization minimizes the redundancies that are present in data tables. Denormalization (reverse of normalization) adds redundant data or group data.