A query optimization and data retrieval system for a high-level database interface
This thesis is to implement query optimization and data retrieval algorithms for a database system based on the Entity-Category-Relationship Model (ECR Model). The ECR Model is an example of a high-level data model in which the burden of efficiency is removed from the user. It becomes the responsibility of the DBMS to automatically optimize the user interface. In this research, a query in the ECR model is translated into a list of operations in an extended relational algebra (RA) before optimization. The query optimization is divided into three sequential phases. They are (1) The transformation of the list to reduce the execution time; (2) Traversing the list downward to label a set of sort orders which can be efficiently generated from the previous operation; and (3) Traversing the list upward to choose the preferred sort order which we want for the next operation and assign a suitable basic procedure for the operation. The data retrieval system retrieves the desirable data from the file system which stores in the database. It takes an operation list and implements each operation from 16 basic procedures in such a way that the sort orders of intermediate relations are optimally coordinated.