Dolování a kategorizace SQL příkazů ze StackOverflow databáze

Abstract

The main topic of this diploma thesis is to find similar SQL queries that are similar to the given input SQL query. This diploma thesis is working with a collection of SQL queries from StackOverflow. StackOverflow is a website that solves programming problems and contains questions and answers related to the source code of various programming languages. In this thesis, the similarity of the SQL query is calculated using the XML similarity metrics, based on the similarity of two XML documents. The similarity metric had to be tested both in terms of correctness and implementation efficiency. The result is a web information system that allows you to search for similar SQL queries to the specified input query and to the selected RDBMS (currently it is PL/SQL, T-SQL, MySQL, and SQLite). The application uses an inverted index and works with better than linear complexity.