Diagnosing faults in embedded queries in database applications

  • Authors:
  • Muhammad Akhter Javid;Suzanne M. Embury

  • Affiliations:
  • University of Manchester;University of Manchester

  • Venue:
  • Proceedings of the 2012 Joint EDBT/ICDT Workshops
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Diagnosing faults in embedded queries in database applications is a daunting process. When test cases fail, the traditional way of diagnosing faults is to follow possible execution paths, either mentally or step-by-step in a debugger, to locate the problematic area. The diagnosis problem becomes even harder when you have embedded language with quite different semantics and properties. Our focus is on a specific problem: diagnosing failed test cases caused by embedded queries in database applications which are syntactically correct but semantically incorrect (i.e., they produce incomplete or incorrect results). Much research literature is available on database applications and databases but the diagnosis problem for embedded queries that cause failure of test cases has not been tackled. We perform an experiment to see how far existing techniques could be useful in proposing a new technique for this problem. We identify the additional components that need to be developed to take us to a full solution and describe our tentative conclusions so far.