Cluster Algebra: A Query Language for Heterogeneous Databases

This report describes a query language based on algebra for heterogeneous databases. The database logic is used as a uniform framework for studying the heterogeneous databases. The data model based on the database logic is referred to as cluster data model in this report. Generalized Structured Query Language (GSQL) is used for expressing ad-hoc queries over the relational, hierarchical and network database uniformly. For the purpose of query optimization, a query language that can express the primitive heterogeneous database operations is required. This report describes such a query language for the clusters (i.e., heterogeneous databases). The cluster algebra consists of (a) generalized relational operations such as selection, union, intersection, difference, semi-join, rename and cross-product; (b) modified relational operations such as normal projection and normal join; and (c) new operations such as normalize, embed, and unembed.