No label defined (Q3974567)

The main result of the paper can be characterized as the contribution to the optimization process of queries accessing relational databases. Relational database systems usually provide high-level languages which specify data to be accessed without describing how it shall be done. Low- level primitive functions perform the actual data access the efficiency of which depends both on the physical organization of data and on the equivalence with the respective high-level sentences. In order to establish a mathematical basis to the optimization techniques a minimum set of relational algebra operations, called \(\gamma\)-algebra, is defined. It consists of three operations, which are \(\gamma\)- projection, \(\gamma\)-selection and \(\gamma\)-union, and has the completeness property. A cost function capable of quantifying the tuple access number is defined. This function makes it possible to evaluate the differences between two equivalent \(\gamma\)-algebra expressions with respect to the point of view of the data access efficiency. Finally, an optimization strategy for \(\gamma\)-algebra expressions is described. The optimization criterion is to reduce the number of tuple accesses which can be quantified by using cost functions for each equivalent query. This reduction has actually been shown through the application of some cost functions previously defined.