As database query language is more frequently used to get useful information from very large database, the development of rather user-friendly query interface has been promoted. End users expect the misconception handling of query statement which is written by lay users, the intelligent analysis of query whose meaning is vague, and the presentation of approximate result for the query that has no exact answer. The Knowledge Abstraction Hierarchy is invented to maximize the user satisfaction through transforming the original query using the knowledge about query and data, and providing approximate result or related result that is meaningful. This paper devises a scheme of Knowledge Abstraction Database on the basis of Knowledge Abstraction Hierarchy, and proposes the KaSQL which is the enhanced query language to support the various operations for maintenance/repair of Knowledge Abstraction Database and retrieval of data. KaSQL analyzes operations contained in input query, maintains the knowledge of KaDB, and processes the query whose meaning is vague, such as approximate selection, approximate join, conceptual selection, and conceptual join through generalization and specialization of original query. The prototype system has been implemented as 3-tier model by laying Java Application at middle-tier.
