Mode Definition And Integration Of Semi-structured Data

It is more difficult to distinguish the application boundary between structured and semi-structured data as the Internet comes into life, the technology of integration between the two has already been an important research subject in fields like Network, Database. Non-pattern and self-description have made semi-structured data easy to use, meanwhile, they also bring difficulty in integration including semi-structured and semi-structured data, semi-structured and structured data. Therefore, the issue needs to address first is pattern definition of semi-structured data.A method of pattern definition and integration of semi-structured data, which is based on structure analysis, was proposed in this thesis, related concepts were defined to realize the definition of pattern, and to establish data model, and that also leads to a unified solution to address issues about data integration. We have studied the pattern, model and integration related theories and technologies, analysis and comparison of existing theories where also have been made. A method to establish source model and target model was offered by analysis structure characteristic, and concepts were defined to describe every aspect of data model. The mapping rules were made by describing every data model, analysis mapping rule between two models, which can help discover the mapping relationship, and generate mapping file. In the end, the implementation of prototype were presented, which includes key technologies and system design. The approach was proven to be feasible and available by experimental data.