Faster Result Retrieval from Health Care Product Sales Data Warehouse Using Materialized Queries
Existing approaches for result retrieval from a Data Warehouse, i.e., Data Cubes and Materialized Views, incur more processing, maintenance and storage cost. For faster retrieval of query results from Data Warehouse, authors suggest storing executed OLAP
- PDF / 293,924 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 36 Downloads / 173 Views
and Jyotika Doshi
Abstract Existing approaches for result retrieval from a Data Warehouse, i.e., Data Cubes and Materialized Views, incur more processing, maintenance and storage cost. For faster retrieval of query results from Data Warehouse, authors suggest storing executed OLAP queries and their results along with metadata in a relational database referred here as Materialized Query Database (MQDB). For stored queries, processing incremental results using Data Marts is faster as compared to using Data Warehouse. Therefore, a significant reduction in query processing time is achieved using MQDB. Authors depict the working of proposed MQDB approach on the sales data of a health care product manufacturing organization by placing Data Warehouse on Centralized and on Cloud Server. Keywords Data Warehouse · Data Marts · OLAP · Materialized queries · Faster query result retrieval
1 Introduction Results of OLAP (Online Analytical Processing) queries are generated by traversing through warehouse data. For frequent queries, Data Warehouse is invoked repeatedly for generating same results. This is quite time consuming. Major approaches for result retrieval from Data Warehouse are Multidimensional Data Cubes [1–6] and Materialized Views [7–12]. Data Cubes and Materialized Views incur more storage, processing and maintenance cost. There exists a trade-off between materializing of the cube and the cost to materialize them [1]. Various techniques are proposed for reducing the materialization cost of Data Cubes [2–6]. Materialized Views are described as derived relations with respect to the base relations that are materialized by storing in database [7]. Authors [8–12] discuss about the view maintenance overhead issues and discuss various techniques to overcome them. S. Chakraborty (B) · J. Doshi GLS University, Ahmedabad, Gujarat, India e-mail: [email protected] © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 V. Bhateja et al. (eds.), Evolution in Computational Intelligence, Advances in Intelligent Systems and Computing 1176, https://doi.org/10.1007/978-981-15-5788-0_1
1
2
S. Chakraborty and J. Doshi
Authors suggest storing executed OLAP queries with their results and metadata information in a relational database referred here as Materialized Query Database (MQDB). Metadata includes timestamp, frequency, threshold, number of records in output, path of result table and path of Data Mart (for processing incremental data). When an OLAP query is executed, it is determined if its synonymous query exists in MQDB. Two queries generating same results are referred as synonymous queries [13]. Thereafter, it is determined if the query requires an incremental update. If no incremental updates are required, then existing results are fetched from MQDB [13]. For queries requiring incremental updates, incremental results are generated using Data Marts as they are faster compared to using Data Warehouse [14]. Final results are derived by combining stored res
Data Loading...
