| To adapt to the continuous development of the enterprise business, the data model ofenterprise needs the relational evolution. While the enterprise applications on the original datamodel for data operation at a certain stage keeps constant. Without appropriate treatment ofthese operations, the application is still in accordance with the traditional data model tooperate, and data manipulations will inevitably lead to wrong results. As the evolution offrequency and time in the data model is difficult to predict, through the application ofsecondary development to adapt to the evolution of the data model is unrealistic. Thus it isnecessary to need a reasonable way to operate data model after the evolution of the dataoperation under the same circumstances which is to ensure that the original application data.To solve these problems, this thesis presents the method of data manipulationreconstruction during model evolution. The reconstruction operation will be done between thedata submission and data implementation, to ensure that the original user application is notmodified by the data operation. Meanwhile, according to the original approach, it canaccomplish the operation for global data model.The main contents are as follows:Firstly, according to the semantic description without standard global data modelevolution, when the global data model evaluates, the user operations cannot adapt to thechange of data model and structure in semantic aspects. To solve the problem which is thewrong operational object, it introduces the semantic methods of data model based on dataelement. Meanwhile, using the semantic tree, it completes the constraints with applicationscene.Secondly, what is the influence on the user operation is researched in the global modeldata evaluating processing. Combined with the semantic description, it introduces thereconstruction methods which is based on model structured features and semantic features. Itcomplete the mission which is operation reconstruction in model evaluation.Thirdly, from the data independence, it analyzes the user data operation. Meanwhile, itintroduces the way that it builds the view that it restores the user data model between userapplication and global data model. And it makes this way the foundation on the reconstructionstrategy based on user data model view. Meanwhile, it guarantees the dependence on the useroperations. Consequently, through the study of data reconstruction, it describes the method which isto achieve the system about data manipulation reconstruction in the procedure of modelevolution, to verify the effectiveness and feasibility of the proposed model in the evolution ofdata reconstruction. |