A query inversion technique for detection of unexpected values in relational databases

Abstract

. Day by day data volumes are increasing, and most of the data are stored in the databases after manual transformations and derivations. The behavior of those stored data is unpredictable. Furthermore, the data are collected from various sources such as physical, geological, environmental, chemical, and biological. A relational database management system (RDBMS) provides a high level data interface. Inside RDBMS sources and intermediate data items are relations, tuples, and attributes. In the context of data provenance, this paper describes how data are produced. When data needs to be retrieved from RDBMS using queries, sometimes it is necessary to check the output data product back to its source values if that particular output seems to have an unexpected value. The aim of this paper is to show the source values for output data using query inversion approach, and to propose the technique for creating an inverse query for queries with aggregation functions, multiple (join, set) operations, and sub-queries.