Processing of multiple processing queries in distributed databases

A set of algorithms is described that can be used to reduce the complexity of evaluating multiple queries of a transaction in a distributed environment. With the consideration of conjunct sharing, it compiles a set of queries into a network based on the concept of semijoins. As some of the queries in a transaction may change the contents of a database, evaluation of the network corresponding to the transaction is synchronized into several phases so that the dependencies among the queries can be properly captured. It is shown how a transaction that includes database updates can be evaluated incrementally in multiple phases such that the states of the evaluation process can be saved and only part of the transaction which is affected by a change needs to be reevaluated. The algorithms described can be applied to relational databases with slight modifications