Query processing in multidatabase systems

An outerjoin operation is important in the multidatabase systems, but the operation produces unnecessary null values. Resolving inconsistencies of data by an aggregation could give users misleading information, cause unnecessary data movement which increases query processing costs, and also violates the autonomy of local databases. An efficient algorithm for solving these problems in multidatabase systems is presented. We generate the answers as a multiple set by which removing the null values. We give users more accurate data information by presenting all data values from different local databases, and give users flexibility by allowing them to define their own query processing strategy for the resolution of the inconsistent data. Our approach also reduces data movement and local processing costs by identifying data which will not be included in the answer set as early as possible, thus reducing data transmission costs and local processing costs. Our principle of query processing is that we allow only necessary data movement between sites so that we can utilize local processing as much as possible and make global site processing simple by having it handle small amounts of data and use simple processing strategies.