Anda di halaman 1dari 7

Define : Confirmed Fact Conformed dimensions are the dimensions which can be used across multiple Data Marts

in combination with multiple facts tables accordingly. Data Mart in Data Warehousing A data mart (DM) is a specialized version of a data warehouse (DW). Like data warehouses, data marts contain a snapshot of operational data that helps business people to strategize based on analyses of past trends and experiences. The key difference is that the creation of a data mart is predicated on a specific, predefined need for a certain grouping and configuration of select data. A data mart configuration emphasizes easy access to relevant information (Reference : Wiki). Data Marts are designed to help manager make strategic decisions about their business. Conformed Dimensions in Data Warehousing Conformed dimensions mean the exact same thing with every possible fact table to which they are joined. They are common to the cubes. Define : ODS (Operation Data Source) ODS - Operational Data Store. ODS Comes between staging area & Data Warehouse. The data is ODS will be at the low level of granularity. Once data was populated in ODS aggregated data will be loaded into into EDW through ODS. It is also a similar small DWH which will help analyst to analysis the business. It will have data for less number of days. Generally it will be around 30-45

days. Like DWH here also we will primary keys will be generated, error and reject handling will be done. Its Conformed Dimensions. To put in simple English they are Dimensions which provide the same meaning irrespective of the Fact Table they are associated with. If you have a Dimension then it can be used in Multiple Data Marts or say Multiple Fact Tables and still provide the same unified view without having to duplicate them. Good Example are Time Dimension, Product & Customer. Conformed Dimensions are nothing but dimensions, but they are linked to many fact tables where as the Dimensions are meant to be linked with its own Fact table. The following figure gives the exact difference between these two...

Here F1, F2,... are Fact tables D1, D2... are Dimension tables where as D3, D4 and D5 are Conformed Dimensions.

A small update on Conformed Dimension from my side Confirmed dimension is a dimension modeling technique promoted by Ralph Kimball. A Conformed Dimension is a dimension that has a single meaning and content throughout a data warehouse. A conformed dimension can be used in any star schema. For Example - Time/Calendar dimension is normally used in all star schemas so can be designed once and used with many fact tables across a data warehouse. The use of conformed dimensions and shared measures is the primary way a set of data marts can be united into one consolidated data warehouse. Conformed dimensions are what which are common between data marts and make the drill across possible. Conformed dimensions Conformed dimensions can be used to analyze facts from two or more data marts. Suppose you have a shipping data mart (telling you what youve shipped to whom and when) and a sales data mart (telling you who has purchased what and when). Both marts require a customer dimension and a time dimension. If theyre the same dimension, then you have conforming dimensions, allowing you to extract and manipulate facts relating to a particular customer from both marts, answering questions such as whether late shipments have affected sales to that customer. Conformed facts In addition to conformed dimensions, you need conformed facts. Conforming a fact needs standardizing the definitions of terms across individual marts. Often, different divisions or departments use the same term in different ways. Does revenue refer to gross

revenue or adjusted revenue? Does units shipped refer to cases of items or individual items?so consider this also. Conform facts and dimensions to evolve your data warehouse over time By Dan Pratte April 30, 2001, 7:00am PDT An enterprise data warehouse can prove essential to the success of your business. But despite its value, building one is a daunting project. It can take years and cost millions. To show faster results at less cost, some experts recommend the alternative approach of constructing smaller data marts, which can later be combined into an enterprisewide data warehouse. (See my previous article,Data marts deliver fast results, but proceed with caution.) This second, bottom-up approach is favored by most enterprises. But to ensure the success of a data mart project, some brief, but careful, enterprisewide collaboration and planning must take place at the inception. Otherwise, the completed data marts will simply replicate the problems they were intended to resolve: data redundancy, islands of information, disparate systems and interfaces, arguments over who has the right numbers, and a host of other troublesome issues. This article outlines the essential up-front planning that enables the creation of functional data marts that will later serve as the components of an enterprisewide data warehouse. The data warehouse bus architecture Initial planning activities should focus on developing and documenting an overall proper architecture to which individual data marts will conform. If you do so, youll realize two key benefits:

Your individual data marts will integrate seamlessly into the fabric of the enterprise data warehouse. Youll realize benefits from the data marts quickly, and the marts will allow you to develop and evolve your enterprise warehouse over time.

Data warehouse bus architecture (DWBA), advocated by data warehouse expert Ralph Kimball, is the best architecture to follow. The term bus in this senseconceptually akin to the expansion-bus in your computer allows you to plug and play data marts like a plug and play PC video card. The enterprise data warehouse consists of separately implemented data marts bound together with a powerful architecture based onconformed dimensions and conformed facts, to quote Kimball. This sounds exactly like what were after, but what are conformed dimensions and facts? Dont miss the previous installments in this series: "The intelligent organization: An introduction to BI": In this article, Pratte talks about what business intelligence can do for an organization and offers some examples of how BI is used within the enterprise. "Use this architecture to structure your business intelligence solutions": Pratte outlines the activities, application types, and data structures necessary to process an enterprise's raw data into useful information. "Building the business intelligence solution: Back-room issues": In this article, Pratte discusses the back-room BI activities, which involve the design and population of the data warehouse and the production of the multidimensional cube. Business intelligence: Finding the best format for your end users: Pratte shows you how to tailor your BI tools to particular types of users so your organization can reap the full benefit from your BI system. Thinking dimensionally aids business intelligence design and use: This article outlines the dimensional data model and how to structure your data warehouse using this architecture. Data marts deliver fast results, but proceed with caution: Building data marts first may be the best choice for some organizations. Pratte gives you some tips to ensure the success of this approach. Conformed dimensions Conformed dimensions can be used to analyze facts from two or more data marts. Suppose you have a shipping data mart (telling you what youve shipped to whom and when) and a sales data mart (telling you who has purchased what and when). Both marts require a customer dimension and a time dimension. If theyre the same dimension, then you have conformingdimensions, allowing you to extract and manipulate facts relating to a particular customer from both marts, answering questions such as whether late shipments have affected sales to that customer.

Suppose now that you add a marketing data mart to help you analyze product promotions. Again, with conformed customer and time dimensions, youre able to analyze the effects of a particular product promotion on sales. (Analyzing facts from more than one fact table in this way is termed drilling across. My previous article, Thinking dimensionally aids business intelligence design and use, explains the function of facts and dimensions.) As this example shows, the very same conformed dimensionsin this case, time and customer dimensionshave meaning in the context of three independentlydeveloped data marts. These dimensions become enterprise property and can be used later in other marts as you evolve the enterprise data warehouse. Conformed facts In addition to conformed dimensions, you need conformed facts. Conforming a fact really amounts to standardizing the definitions of terms across individual marts. Often, different divisions or departments use the same term in different ways. Does revenue refer to gross revenue or adjusted revenue? Does units shipped refer to cases of items or individual items? Make certain your design team develops, early on, a uniform enterprise taxonomyand enforce it. When marts can stay marts Sometimes you dont need to be so concerned with an architecture aimed at tightly integrating your marts, and a particular mart can just stand alone. Data marts constructed to focus analytical support to solve a single business problem are examples of this and are termed disposable data marts. Heres an example: A transporter of new automobiles was bearing the burden of costly vehicle damage claims. A data mart was built in order to gain insights into the causes for the claimswhich routes were taken on which days and which parts of the cars were being damaged. As a result of the analysis, procedural changes reduced damage claims to acceptable levels and the mart was then discarded. Conclusion

Remember these three key points when planning a bottom-up approach to data mart construction:

The bottom-up approach enables you to quickly realize the benefits of lowcost and easy-to-build data marts while the enterprise data warehouse evolves over time by combining the individual marts. Using conformed dimensions and facts in your independent data marts allows you to combine them later. If you fail to conform dimensions and facts, your marts will be stovepipe marts, and youll have little hope of integration. Youll wind up with data islands and disconnected information. The data warehouse bus architecture forces compliance on these points. Have you undertaken an enterprise data warehouse initiative? Does your experience include addressing some of the questions and issues discussed in this article? Share your thoughts.