Query Formulation and Recommendation for Relational Databases Using User Sessions and Collaborative Filtering

Abstract

Structured Query Language (SQL) has a uniform structure over different programming languages. The queries fired on Database Management System (DBMS) contain textual information along with selected segments of data parsed by data base management system to fire it as a structured query. Currently DBA needs to execute complex queries on large databases. Many times user or DBA fires similar queries on database server to get useful information. The queries which are similar to each other can then be categorized into two types a) the tuples retrieved by SQL queries are similar b) the fragment of the queries are similar. System gives recommendation to those similar queries so that it saves the time of DBA to construct it again and again. Query suggestions given to DBA or users are known as Query Recommendation. To develop a Query Recommendation system many authors suggested the use of Query Log. Query suggestions are divided into two areas mainly Collaborative Recommendations and Single Log Recommendations. This system is designed by single or collaborative log using parameter known as mixing factor. In this paper we analyzed Sql query Recommendation concepts and their uses. There are basically two types of similarity measure for Query Recommendation considered in [1] such as 1) Fragment Based 2) Tuple Based. Here in this research paper we are motivated towards generating recommendations for nested SQL queries. We adopt hierarchical classification on query log to create classes of similar queries and further to generate recommendations for SQL Query we proceed with finding matching class from which the recommendations can be modeled.
DOI: 10.17762/ijritcc2321-8169.150704