Research On Data Migration Algorithm Based On Heterogeneous Table Structures

With the development of computer technology and the advent of the Internet age, theupgrading of the application system speeds up data as the core of the application system, thesame faces under the migration problem in a different system and environment. Most of thedata migration needs exist in the relational database, in order to solve the problem ofheterogeneous data in relational database, many research institutions and scholars lot of effortfrom all aspects, and propose a solution for the mutual transformation of heterogeneous data,so that data can be accurate, efficient migration between different databases. In many researchdirection, the table structure heterogeneous problem is one of the more difficult migrationbarriers.This paper focused on the heterogeneous table data migration issues and mainstreamalgorithms, including variable length chromosome genetic algorithm and the minimumspanning tree algorithm. Based on the lack of genetic algorithm for multi-node processing,and the minimum spanning tree algorithm derived tables’ stability of the operating efficiencyof defects, this paper found the idea of its expansion layer application of certain performanceadvantages in the derivative of the graph structure of the junction by anglicizing Dijkstraalgorithm.This paper modified the ER data model and added derived table node to rebuild thegraph structure, and then converted the multi-node in the graph structure to a couple oftwo-node problems, so that the new data structure could be applied to the efficient algorithm,which improved in accuracy, stability and efficiency. In order to solve the heterogeneous tabledata migration problem, this paper not only designed key data structure but also completedthe rest parts of SQL generator, and then got a practical and executable solution. The solutionwas implemented with Java language. Experiments were taken to verify this solution, inputinformation was parsed and processed, and then the transformation was completedautomatically by using SQL. Finally, it proved this algorithm had certain advantages over thetwo mainstream algorithms.