Persistent Errors in Query Formulation

Abstract

We use the internet daily to query data from a myriad of databases; every search term entered in a search engine, every movie watched, every song listened, ev-ery newspaper article read online. Although we as end-users only see the rel-atively effortless user interfaces as we query data, someone has had to formal-ize our queries into a language the software understands. The most common of these so called query languages is Structured Query Language (SQL). In or-der for us as end-users to retrieve exactly the data we want, it is crucial that the software developers responsible for writing the underlying queries have written the queries without errors. Educational SQL research, however, has not yet thor-oughly addressed issues related to understanding query formulation errors or some technical factors which influence the process of learning SQL. This doctoral dissertation makes the following contributions for increased understanding of SQL education: (i) a systematic overview of SQL teaching practices proposed in scientific literature, (ii) a creation of a wide taxonomy of errors committed in SQL learning, (iii) a description of which types of errors halt query formulation, and which types of encountered errors are usually fixed, (iv) evidence on the effects of database complexity on query formulation success rates, and (v) a creation of a planning notation designed to mitigate errors in query formulation. Contribu-tion (ii) presents practical implications for research by allowing the comparison of results of different SQL error studies when the taxonomy is used, and extend-ing and generalizing prior SQL error studies. While contributions (i) and (v) may be directly applied in teaching SQL, contributions (iii) and (iv) may be consid-ered when making an informed decision on what kind of databases are the most suitable for practicing SQL.