The Research And Implementation Of Data Source Change Detection Methods Based On Snapshot Differential Algorithms

In order to solve the extensive phenomenon of "detached island of information",heterogeneous data integration has become one of the hotspots in the present research.In this paper, we study and realize the first step of data integration, which is monitorto information source. We mainly study the change detecting method based on thesnapshot differential technology, and our application support various kinds ofcommercial relation databases as information sources.Firstly, we introduce the Data Warehouse architecture of the WHIPS project ofthe Stanford. Then, we introduce the design of the Change Detection Module, and weintroduce the procedure based on analyzing log records and which based on snapshotdifferential technology.Secondly, we describe several snapshot differential algorithms. We comprehendand master their application scope, processing speed and I/O consumption byanalyzing in theory and in experiment. Then, we focus on introducing windowalgorithms, including its excellence, faults and improvement.At last, we introduce some technologies in detail which involved in realizing oursnapshot differential module, including view definition, snapshot creation, snapshotdifference creation.