Intelligent Application of Partial Repair for Handling Inconsistency among Database

Abstract

Handling inconsistencies among standalone and integrated databases has been an important issue for database’s administrators for decades where nowadays databases are huge and not only different types of inconsistencies are ubiquitous in them but also application of any repair might induce new violations to integrity constraints. Resolved data may harm rest of database that leads to a costly process for repair of inconsistent data while after any resolution of data, database should be checked whether any new violation has emerged or not. Introducing partial repair through an approach to measure the tendency that a resolved portion of data incurs new violation would help any repair algorithm to isolate a selection of problematic data (not all), resolve it and save the database from being hurt during repair process. Partial repair keeps the rest of data from being affected that eliminates concerns over application of repair. Partial repair may not handle entire inconsistencies among databases but it represents a repair that would have minimum harm to rest of data along with consideration of cost which makes it valuable.

Keywords: data quality, repair, inconsistency, dependency