A Framework for Computing the Overlap and Difference of Relational Schemas

Mohammad Sadoghi.

Technical Report CSRG-609 (MSc), University of Toronto, 2008.

Abstract

This thesis focuses on developing the Overlap and Difference operators for managing the relational model in a data exchange setting. A data exchange setting consists of a source schema and a target schema along with a mapping that transforms a source instance into a target instance. We define the overlap and the diff settings and provide algorithms to implement the Overlap and the Difference operations for relational schemas. An enterprise system that operates on multiple heterogeneous database systems often is required to manage independently created relational data sources by creating a unified schema of the data, a merged schema; finding a schema of common elements within two schemas, an overlap schema; or generating schema of elements in one schema that is not represented in another schema, a difference schema. These tasks are labour-intensive and our goal is to simplify the creation of overlap and difference schemas by proposing the Overlap and Difference operators, thereby helping to ease the process of maintaining and building enterprise solutions over large data sources.

Download


Readers who enjoyed the above work, may also like the following:


  • Optimizing Key-Value Stores for Hybrid Storage Architectures.
    Prashanth Menon, Tilmann Rabl, Mohammad Sadoghi, and Hans-Arno Jacobsen.
    In Proceedings of CASCON, 2014.
    Tags: key-value stores, leveldb
  • Adaptive Parallel Compressed Event Matching.
    Mohammad Sadoghi and Hans-Arno Jacobsen.
    In 30th IEEE International Conference on Data Engineering, 2014.
    Tags: content-based matching, publish/subscribe, event processing
  • CaSSanDra: An SSD Boosted Key-Value Store.
    Prashanth Menon, Tilmann Rabl, Mohammad Sadoghi, and Hans-Arno Jacobsen.
    In 30th IEEE International Conference on Data Engineering, pages 1162-1167, 2014.
    Tags: cassandra, big data, key-value store, nosql