Evaluation of different designs to represent missing information in SQL databases

Abstract

It is possible to use different designs to deal with the missing information problem in SQL databases. In this paper, we use a multi-criteria decision support method, which is based on the Analytic Hierarchy Process, to evaluate a set of possible designs for dealing with missing information. One of the designs uses SQL NULL-marks, another design uses tables that are in the sixth normal form, and one of the designs uses special values to represent missing information. We evaluate the designs, which are implemented in a PostgreSQL™ database, in terms of two hypothetical contexts. We create a test database, perform measurements, and use the results to compare the designs. The results of the evaluation provide new insights into the advantages and disadvantages of the different designs. © 2013 Springer Science+Business Media.