Temporal and Contextual Dependencies in Relational Data Modeling

Although a solid theoretical foundation of relational data modeling has existed for decades, critical reassessment from temporal requirements’ perspective reveals shortcomings in its integrity constraints. We identify the need for this work by discussing how existing relational databases fail to ensure correctness of data when the data to be stored is time sensitive. The analysis presented in this work becomes particularly important in present times where, because of relational databases’ inadequacy to cater to all the requirements, new forms of database systems such as temporal databases, active databases, real time databases, and NoSQL (non-relational) databases have been introduced. In relational databases, temporal requirements have been dealt with either at application level using scripts or through manual assistance, but no attempts have been made to address them at design level. These requirements are the ones that need changing metadata as the time progresses, which remains unsupported by Relational Database Management System (RDBMS) to date. Starting with shortcomings of data, entity, and referential integrity in relational data modeling, we propose a new form of integrity that works at a more detailed level of granularity. We also present several important concepts including temporal dependency, contextual dependency, and cell level integrity. We then introduce cellular-constraints to implement the proposed integrity and dependencies, and also how they can be incorporated into the relational data model to enable RDBMS to handle temporal requirements in future. Overall, we provide a formal description to address the temporal requirements’ problem in relational data model, and design a framework for solving this problem. We have supplemented our proposition using examples, experiments and results