Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

174 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

5.1 Temporal data management concepts and use

cases

Considerations regarding time have been part of the concepts of data

warehousing and business intelligence from their beginnings. It is hardly possible

to conceive of business data or business queries that do not contain an element

of time, either as the subject (“when” did something happen) or the context (tell

me what happened during a certain time period) of the query or data. Even when

the value of time is not explicit, for example, in a list of customers, it is still

implicit, as in a list of customers “at this point in time.”

Events and transactions of interest to a business or enterprise always occur at a

particular time (orders are placed, currencies change hands, applications are

approved, warranties expire); they all happen at a moment in time that is

recorded in some transactional system and ultimately reflected in the data

warehouse.

Knowing when events occurred allows the enterprise to ask an endless number

of questions and perform various analyses to understand how the business is

performing, how metrics are trending relative to comparisons across time

periods, how business performance might change in the future, and so on.

Indeed, achieving meaningful business intelligence and reporting is not possible

without the inclusion of time in the data warehouse and subsequent analyses.

This is undoubtedly apparent to those who have experience with data warehouse

or business intelligence applications (or who have ever managed their own

checkbooks). And, because data warehousing technology has been around for

decades, it might seem likely that database management systems built for data

warehousing provides built-in features to manage notions of time.

However, except for a few temporal data types and time-based functions, this

has largely not been the case. Data warehouse architects, database

administrators (DBAs), and application developers have generally been left to

construct their own models for representing time and analyzing information

relative to time. Although several industry “best practices” have arisen over the

years that provide consistent techniques to solve common problems, for the

most part there has been a vast proliferation of custom and ad hoc solutions to

managing time in a data warehouse. Until recently, that is.

Now, IBM InfoSphere Warehouse 10 includes for the first time native temporal

awareness and management features in DB2 that place IBM in the forefront of

data warehouse technology providers. In this chapter, we explore how this

feature works and how architects, DBAs, and application developers can use it to

readily support any number of time-based use cases.