Introduction to Query Processing

Query processing in databases can be divided into two steps: selecting an 'optimal' evaluation strategy, and executing it. We first present elementary nested loop and relational algebra algorithms for query execution and point out some opportunities for improving their performance. A survey of optimization strategies, structured in query transformation techniques and access planning methods, follows. Finally, extensions for special-purpose query systems are briefly addressed. 1.0 PERFORMANCE CONSIDERATIONS IN DATABASE SYSTEMS Database management systems (DBMS) are now a widely accepted tool for reducing the problem of managing a large collection of shared data for application programmers and end users. The user interacts with the system by submitting requests for data selection (queries) or manipulation (updates). Both kinds of operations frequently involve access to data described to the system in terms of their properties rather than their location. A sequence of queries or updates which is logically a single unit of interaction with the database is called a transaction. To fulfill its mission, a DBMS must be efficient in the sense that it minimizes the consumption of human and machine resources for processing transactions submitted to it. The costs of human resources in utilizing a DBMS are determined, among other factors, by the power and friendliness of the language provided to each type of user (application programmer or end user), and by the system's response time. The goals of language power and fast response time may be in conflict since it is often difficult to implement a powerful language construct efficiently. It is the task of the database implementor to reduce this potential problem. Machine resources used by the DBMS include the storage space for data and access paths in secondary memory, as well as for main memory buffers, and the time spent by the CPU and channels for data transfer to and from secondary memory and other computers (in distributed databases). The trade-off between these cost components is influenced by the architecture of the database system. This work was supported in part by the Deutsche Forschungsgemeinschaft under grant no. SCHM 450/2-1. Center for Digital Economy Research Stem School of Business IVorking Paper IS-84-48