Book Image

Oracle Data Integrator 11g Cookbook

Book Image

Oracle Data Integrator 11g Cookbook

Overview of this book

Oracle Data Integrator (ODI) is Oracle's strategic data integration platform for high-speed data transformation and movement between different systems. From high-volume batches, to SOA-enabled data services, to trickle operations, ODI is a cutting-edge platform that offers heterogeneous connectivity, enterprise-level deployment, and strong administrative, diagnostic, and management capabilities."Oracle Data Integrator 11g Cookbook" will take you on a journey past your first steps with ODI to a new level of proficiency, lifting the cover on many of the internals of the product to help you better leverage the most advanced features.The first part of this book will focus on the administrative tasks required for a successful deployment, moving on to showing you how to best leverage Knowledge Modules with explanations of their internals and focus on specific examples. Next we will look into some advanced coding techniques for interfaces, packages, models, and a focus on XML. Finally the book will lift the cover on web services as well as the ODI SDK, along with additional advanced techniques that may be unknown to many users.Throughout "Oracle Data Integrator 11g Cookbook", the authors convey real-world advice and best practices learned from their extensive hands-on experience.
Table of Contents (19 chapters)
Oracle Data Integrator 11g Cookbook
Credits
Foreword
About the Authors
About the Reviewers
www.PacktPub.com
Preface
Index

Loading data using partition exchange


Table partitioning is a commonly used strategy in data warehouse systems, and although the many reasons for partitioning a table are beyond the scope of this book, suffice it to say that performance and high availability are among the primary reasons for doing so. From a data loading perspective, the gold standard ELT strategy for targeting a partitioned table is universally considered to be the partition exchange loading (PEL) method. This advanced technique involves five basic steps:

  1. Create a standalone database table that is structurally identical to the partitioned table.

  2. Load all source records into the standalone table that comply with the boundary definition of a specific partition within the partitioned table.

  3. Create all secondary objects (local indexes, constraints) for the standalone table such that the final structure is identical to that of the partitioned table.

  4. Exchange the contents of the standalone table with that of its corresponding partition...