Method of integrating in real time large volumes of updates in a database

摘要

A method of integrating large volumes of updates in a database system is described. Each individual update is comprised of a coherent set of data records supplied by a data provider. The database system includes a master file repository and an active image repository. The method comprises the steps of first obtaining for each individual update a unique modification identifier from a logistic table of the database system. The master file repository is then updated with the coherent set of data records of each uniquely identified individual update. A unique commit identifier is further obtained from the logistic table for each individual update that has been committed by the master file. The active image repository is synchronized by successively loading into the active image, in the order specified by the commit identifier, each individual update. Synchronization includes retrieving from the master file repository the coherent set of data records of each uniquely identified individual update. Finally, completion of the propagation of each individual update into the active image repository is achieved when the loading of each corresponding set of data records is committed by the active image. This allows end-users of the database system to start querying immediately the propagated updates from the active image repository.