Improve Data Warehouse Performance by Preprocessing and Avoidance of Complex Resource Intensive Calculations

A Data Warehouse is a computer system designed for archiving and analyzing an organization's historical data, such as sales, customers, products, salaries, or other information from day-to-day operations OLTP. Normally, an organization summarizes and copies information from its operational systems to the data warehouse on a regular schedule, such as daily, weekly, monthly, quarterly or annually; after that, management can perform complex queries and analysis OLAP on the information without slowing down the operational systems. Materialized views can be one best option in this regard and can be used in a number of ways. It can be used in distributed databases for replication and can also be used for efficient provision of data to a query through query re-writing. The process of data provision to queries can further be expedited if dependent child views are created on an already existing materialized view. Furthermore, these child-views are automatically created upon the creation of the base materialized view with some restrictions. This results in less-user dependent activity of creation of materialized views based on some parameters. These parameters are the number of child-materialized views and the type of the data a view contain. In this paper, a balanced approach is suggested to create sub-materialized views to answer user queries without consulting the fact table or parent materialized view that results in avoidance of resource intensive calculations and joining of multiple tables.