Handling Null Values in SQL Queries on Relational Databases

Abstract

The lack of information within relational databases expressed as NULL values presents a problem in ensuring the quality of data and the information retrieved by SQL queries. This is so because there are different interpretations of NULL values. Actually, in many cases, queries either do not consider that NULL values can appear as a result of their evaluation or do not reflect in the code the correct interpretation of these values. NULL values have been present in databases since almost the earliest implementations of relational database management systems, but the implementation of the SQL standard raises multiple problems. When querying on a database that handles NULL values, the results may not be as expected, either by omission of results (false negatives) or by incorrect results (false positives). In this work we propose a methodology that analyses different SQL queries and allows the user to detect possible errors in those queries that have NULL values.