Errors and Complications in SQL Query Formulation

Abstract

SQL is taught in almost all university level database courses, yet SQL has received relatively little attention in educational research. In this study, we present a database management system independent categorization of SQL query errors that students make in an introductory database course. We base the categorization on previous literature, present a class of logical errors that has not been studied in detail, and review and complement these findings by analyzing over 33,000 SQL queries submitted by students. Our analysis verifies error findings presented in previous literature and reveals new types of errors, namely logical errors recurring in similar manners among different students. We present a listing of fundamental SQL query concepts we have identified and based our exercises on, a categorization of different errors and complications, and an operational model for designing SQL exercises.

Crowdsourcing Content Creation for SQL Practice

“…SQL as a language and the formulation of SQL queries can be complex for novices. Several studies have reported about the types of mistakes and errors students encounter when learning to write SQL [3,4,9,37,41,42]. Performance in writing SQL is also linked with students' course performance [2].…”

SQL Education

“…Alternatively, the environment may be a DBMS's SQL interface embedded in a web page, and the web page fitted with auxiliary elements, for example, a representation of the underlying database schema, a natural language request (i.e., data demand) to which a students must write an SQL equivalent, and the correct result table [101]. The database may be represented at the conceptual level as an Entity-Relationship (ER) diagram [37], or at the logical level as a database schema diagram [43].…”

“…Furthermore, semantic errors have been further divided into errors which are evident without knowledge of the underlying data demand, and errors which are only recognizable if the data demand is known [17]. The former kind of semantic errors are called semantic, and the latter logical [101].…”

“…Taipalus et al [101] present a query concept framework based on their teaching experience, an error categorization framework based on a qualitative study, and opinions on how SQL should be taught. By these three aspects, their study could be classified as a philosophical paper, evaluation research, or an opinion paper.…”

“…A recurring theme in the primary studies, regardless of the research topic, was argumentation for [5,8,101,121] and against [75,88,101] natural learning environments. A natural learning environment better reflects industry, i.e., students' future work environments.…”