ATLaS: A Native Extension of SQL for Data Mining and Stream Computations
Haixun Wang · 2002
A lack of power and extensibility in their query languages has seriously limited the generality of DBMSs and hampered their ability to support new application domains. Considerable efforts by database researchers and commercial DBMS vendors have led to major extensions; yet there remain important applications---particularly data mining---that are not supported well in SQL-3. Thus, there is a pressing need for more general mechanisms for extending SQL and dealing with new application areas, particularly database-centric data mining. To satisfy this need, we allow database users to add new table functions and stream-oriented aggregate functions by defining them in SQL---rather than in external procedural languages as O-R DBMSs currently do. This simple extension turns SQL into a powerful database language, which can express a wide range of applications, including recursive queries, ROLAP aggregates, time-series queries, stream-oriented processing, and data mining functions. In addition to adding great power and flexibility to SQL, these extensions are conducive to performance and data independence. The paper also discusses the system we have developed to support these SQL extensions, and the architecture and techniques used in its realization.