Study On SQL Program Structure Similarity Matching And Automatic Assessment Method

Today database technology is a basic means of information processing. The database subject is a required course of domestic university student majored in computer science. The building of the course will have a direct impact on the level of students' quality. Therefore, this paper introduced a scientific, quantitative assessment method into the evaluation of the database course experiment to assess accurately the programming ability of each student, to find problems and improve the teaching process. It has an important significance and practical value on promoting teaching through experimenting and forming a virtuous circle between classroom teaching and extracurricular experiment.Based on principles of procedural similarity measurement methods, this paper made use of structure similar matching method to count the similarity of student's programs, which was seen as the quantitative basis for performance evaluation. The evaluation model was built by combining the static and dynamic strategies. The quantitative evaluation results were given by combining various kinds of metrics. In the matching process of program structure similarity, knowledge was timely pushed by confining different sub-trees. This would provide timely help to solve the problems appearing during the experiments.The main work of this paper is as follows:1. By comparatively analyzing the application scope, the advantages and disadvantages of existing dynamic and static assessment methods, this paper proposed static-based assessment of a combination of both methods, whose static strategies was achieved by program structure Similarity matching and dynamic strategies was implemented by analyzing the program execution results.2. By standardizing the SQL and borrowing the general process and output of the lexical analysis and parsing from Compilers Principles, this paper firstly extracted the abstract syntax tree of SQL. Then with respect to the structural features of the SQL language, the SQL syntax tree, as the similarity measure objects, was gainned by converting the problem domain.3. According to the SQL syntax tree extracted from the program, this paper calculated the similarity between the current student program structure and the optimal model program structure with tree edit distance. Against different nodes in the tree having different effects on the similarity, this paper gave the improved algorithm cost model. At the same time, the effects of similar sub-structure on the overall similarity were measured by defining the similarity factor.4. By adopting the similarity value as the main measure of the static assessment, combined with dynamic assessment, the assessment model framework was designed and the model assessment mechanism was also determined. Analyzing the mapping and conversion table gained from tree matching, algorithms for error location and error classification from structural aspects were designed.5. The prototype system of experimental database platform was designed and implemented in this paper. The whole experimental procedure was simulated by process submission, process evaluation, knowledge mining and other foundational modules. The accuracy of the assessment methods for similarity matching was also verified in the paper.