An Efficient Methodology for the Evaluations of Select Operation in Query Optimization

Abstract

The advancement of 4th Generation Languages (4GLs) brought the mutiny in the Information technology by making users free from the burden of developing logic for solution of the application problem. At the same time, it enhanced the need for implementation of ameliorated strategies to evaluate the intermediate operations. The development of 4GLs freed the users from the strained design of 'how to do it' for solutions of their applications. At the back end, it increased the complexity of transformation from non-procedural query to procedural query. The query processing and optimization manages these transformations, by transforming 4GLs query into appropriate relational algebra/calculus expression in the first phase, minimizing the operations in the next phase and then evaluating the expression through the efficient strategies for such operations in the third phase. The transformation to relational algebra/calculus expression and minimizing operations are straightforward processes. The development of ameliorated strategies depends on the file organizations used in the storing relations of the database (Secondary storage structures). This paper discusses the strategies designed for select operation, when tuples of operand relations are stored in one of various file organizations in the database. Here, an attempt is made to design ameliorated algorithms to evaluate select operation when the tuples of the operand relations are stored in any of the file organizations like (i) sequential file (ii) indexed sequential files with B-tree file having number of key field indexes in each node and B+ tree file organizations for multiple indexes where data pointers are stored in each leaf node of the multilevel indexes of the file organization. The literature available suggests that, there are good methods available when key fields are primary keys. The authors have not noticed any clear cut strategies, when key fields for indexing are non primary keys (non candidate keys). Here, an attempt is made to develop the strategies, when key fields are non-candidate keys.