Mastering Enterprise BI: Time Intelligence Pt. I
February 12, 2007
About the Series ...
This article is a member of the series Introduction to MSSQL Server Analysis Services. The series is designed to provide hands-on application of the fundamentals of MS SQL Server Analysis Services (Analysis Services), with each installment progressively presenting features and techniques designed to meet specific real-world needs. For more information on the series, please see my initial article, Creating Our First Cube. For the software components, samples and tools needed to complete the hands-on portion of this article, see Usage-Based Optimization in Analysis Services 2005, another article within this series.
About the Mastering Enterprise BI Articles...
The purpose of the Mastering Enterprise BI subset of my Introduction to MSSQL Server Analysis Services series is to focus on techniques for implementing features in Analysis Services that parallel or outstrip - those found in the more mature enterprise OLAP packages. In many cases, which I try to outline in my articles at appropriate junctures, the functionality of the OLAP solutions within well-established, but expensive, packages, such as Cognos PowerPlay Transformer and Cognos PowerPlay, can be met often exceeded in most respects by the Analysis Services / Reporting Services combination at a tiny fraction of the cost.
The vacuum of documentation comparing components of the integrated Microsoft BI solution to their counterparts among other enterprise BI vendors, to date, represents a serious undersell of both Analysis Services and Reporting Services, particularly from an OLAP reporting perspective. I hope, within the context of the Mastering Enterprise BI articles, to demonstrate that the ease of replicating popular enterprise BI features in Analysis Services will be yet another reason that the integrated Microsoft solution will commoditize business intelligence.
For more information about the Mastering Enterprise BI articles, see the section entitled About the Mastering Enterprise BI Articles in my article Relative Time Periods in an Analysis Services Cube, Part I.
As I stated in my article Introduction to SQL Server 2000 Analysis Services: Handling Time Dimensions, it is a rare thing to encounter an instance of an OLAP cube that does not require a Time dimension. Throughout years of business intelligence consulting, I have only witnessed this scenario a handful of times within a production environment. Although there often seems to be no shortage of people to argue any side of any statement, few of us would disagree that the measurement of activity over time - and, hence, the Time dimension that supports this capability - is important to both analysis and operational management in general.
As an aside, I refer to the dimension as a Time dimension because my preference is to name dimensions after the generic concepts they represent thus time versus date. While I can certainly live with date as the name of the dimension that represents the concept of time, I do not agree with the argument advanced by some that date is the more appropriate choice because, after all, we are working with date hierarchies that may not subanalyze to time - as in time of day. My response is that date itself is a subordinate member within the larger concept of time, and typically a level within the Time dimension, hence my choice of Time as a dimension name.
(I hope that not too much angst is aroused by Microsofts decision to use terms like time intelligence, server time dimension, time periods, and the like, throughout Analysis Services and its documentation, for those who might confuse time to mean time of day .... Moreover, I heartily encourage substituting date for time as a dimension name when the latter leads to undue stress, justified or not. This is one of the beauties of working within semantic layers...)
The Time dimension has several unique characteristics, relative to other dimensions within our cube models. Among these is the fact that all businesses employ the same core calendar time hierarchy of days (and sometimes lower levels), weeks, and months, together with quarters and years (with various subdivisions included to meet local business and reporting needs) - even though treatment of these various levels can vary widely within the alternative considerations of fiscal years and periods. Moreover, the pervasive nature of time - within and surrounding all organizational activity - means the universal juxtaposition of the Time dimension and the other dimensions within our cube models. Another characteristic of time is its incremental continuation, like a ray in geometry, from a fixed beginning to a typically indefinite end.
The Time dimension has received special focus within the design of enterprise business intelligence applications. Common features include capabilities ranging from the recognition of date fields with minimal intervention to the automatic generation of members of the Calendar time dimension as a part of cube design and / or creation. Most of the dominant applications have even offered support for the dynamic creation of various relative time periods and aggregations. (For a discussion of some of the specific support provided by leading applications, as well as the Analysis Services 2005 approach to meeting and exceeding these features, see other articles within my Introduction to SQL Server 2000 Analysis Services series here at Database Journal).
As we shall see in this article and its successor, Analysis Services 2005 witnesses further enhancements with regard to supporting the Time dimension. Moreover, in addition to these extended features, support for the creation of virtually any custom relative time aggregation that we might need remains readily available to assist developers. In this article, we will gain some familiarity with creating a Time dimension within Analysis Services 2005, focusing upon enhanced features as we encounter them. In Part II, we will examine new features that support the easy addition of Time intelligence within our cube models. Our examination of working with the Time dimension within Analysis Services 2005 includes: