AN ANALYSIS OF DISTRIBUTED DATABASE INDEXING METHOD IN REGARD TO PERFORMANCE OF EXTRACT/TRANSFORM/LOAD (ETL) PROCESSES

Abstract

Data mining and very large databases offer a great deal of promise for business and government alike. Unfortunately, due to the ever changing data needs of the modern organization, as evidenced by the growing number of petabyte data applications, the task to efficiently extract data in a timely manner becomes more difficult. The research described herein presents some encouraging findings in regard to using indexing to improve upon the ETL process. Specifically, a decrease of about 3 times in run time was observed when moving from a flat table to an indexed table function logic. The business value of this decrease may be realized in terms of less wait time when end-user queries are executed.