Reporting Principles

Reporting is an oft overlooked element within both business and systems analysis. Many system designs assume that the database designed to support transaction processing, or worse, the direct implementation of the logical data model will meet reporting needs. While there is some acknowledgement of the need to separate transaction processing from on line reporting with the development of data warehouses, the principles and the motivations behind this separation are often not fully understood. The result is quite frequently systems that cannot meet the business needs because of the clash between reporting and transaction activity.

Transaction Needs	Reporting Needs
Indexes on all tables minimized to reduce transaction write time	Indexes maximized to allow for unexpected requirements for specification of reports, chiefly selection criteria
Validation Tables maintained as separate data structures to allow for flexibility in transaction processing. (Particularly in setting where multiple languages are required or business requirements evolve or or not fully understood)	The number of tables required to produce a report need minimized to meet report production goals
Calculation performed to support validation requirements	Calculations performed to support operational, strategic and performance needs

Compounding this, making a habit of producing large complex reports from a data base also supporting heavy transaction activity means that one or both activities will suffer

Developing a set of principles therefore that recognize these needs and constraints to build a suitable system design that will accommodate a specific set of requirements is critical to the success, no some much of a project, but of the final product as a useful tool for the business.

[ Home ] [ Up ] [ Requirements ] [ Reporting Principles ] [ Data Warehousing ]

Copyright © Neil Kemp and Associates Ltd