Experience Report on Using LANTERN in Teaching Relational Query Processing

The database systems course in undergraduate computer science programs has become increasingly important due to the widespread use of relational databases (RDBMS) and the growing field of Data Science. A primary learning objective for students in this course is to understand how SQL queries are executed in practice within an RDBMS. Typically, an off-the-shelf RDBMS provides a query execution plan (QEP) in either visual or textual format, detailing the steps taken to execute a given query. However, these QEPs contain vendor-specific implementation details that can often be overwhelming for students to grasp. In this paper, we discuss our experience of using a state-of-the-art tool called lantern in teaching QEP s to two cohorts of students enrolled in the database systems course at Nanyang Technological University. Specifically, lantern generates a natural language (NL)-based description of the execution strategy (QEP) chosen by the underlying RDBMS to process a user-specified query. We emphasize on how lantern serves as a supplementary tool that allows students to explore and learn about QEPs related to their queries. Additionally, we analyze the correlation between students' academic performance and their usage of lantern. Drawing on our experiences, we discuss future directions of lantern -augmented learning of relational query processing.