Book Image

Oracle Business Intelligence 11g R1 Cookbook

By : Cuneyt Yilmaz
Book Image

Oracle Business Intelligence 11g R1 Cookbook

By: Cuneyt Yilmaz

Overview of this book

<p>Extracting meaningful and valuable business information from transactional databases is crucial for any organization. OBIEE 11g is a reporting tool that satisfies all the business requirements regarding complex reporting. It consists of a powerful back-end engine with a repository and a highly customizable graphical web interface.</p> <p>Oracle Business Intelligence 11g R1 Cookbook provides all the key concepts of the product including the architecture of the BI Server. This practical guide shows each and every step of creating analytical reports starting from building a well-designed repository. You will learn how to create analytical reports that will support different business perspectives. <br /><br />This practical guide covers how to implement OBIEE 11g suite in order to enable BI developers to create sophisticated web based reports. All of tasks will be covered step by step in detail. <br /><br />You will explore the architecture of the Oracle Business Intelligence Server and learn how to build the repository (RPD). We will also discuss how to implement the business rules in the repository with real-life scenarios.</p> <p>Best practices of a successful BI implementation are esssential for any BI developer so they are also covered in depth.If you are planning to implement OBIEE 11g suite, this step-by-step guide is a must have resource.All the key tasks are defined in detail and supported with diagrams and screenshots.</p>
Table of Contents (19 chapters)
Oracle Business Intelligence 11g R1 Cookbook
Credits
About the Author
About the Reviewers
www.PacktPub.com
Preface
Index

Building the Business Model and Mapping layer


The Business Model and Mapping layer defines the logical model of the data that maps the business rules to the Physical layer objects. There can be many business models that can be mapped with different kinds of data sources based on different technologies. Also dimensional modeling should be done in this layer. This can be achieved by creating dimension objects and their hierarchies. Also all about the calculation and aggregation rules will be defined in this layer.

As a summary, all the business rules are going to be created in this layer. It's probably the most important layer in the repository, and it'll definitely take longer to create the objects in this layer.

Creating business models can be done by two methods, manually or by drag-and-drop. If the manual object creation method is preferred, we'll have to create objects in the following order:

  1. Create the business model.

  2. Create the logical tables and map them to the physical tables.

  3. Create the logical columns.

  4. Define the logical joins.

How to do it...

  1. Create the business model manually by right-clicking on the BMM layer and selecting New Business Model. (Give it a name, in this case Sales.)

  2. Ensure that the Disabled option is checked because there's no other object in this business model. We're going to change this setting later.

  3. After creating a blank and disabled business model, drag-and-drop the physical objects from the Physical layer onto the business model named Sales.

    Automatically created objects are:

    • Logical dimension and fact tables

    • Logical columns and their mappings to Physical layer

    • Logical table sources for the dimension and fact tables

    • Logical keys and joins (can be accessed from the Business Model Diagram window)

  4. Now it'll be good to make a clean-up because with this automated process all of the columns from the Physical layer are created as logical columns. Most probably we won't need all of the logical columns; we can easily delete the logical columns that are not required by right-clicking on the columns. Another task is renaming the logical tables and columns with meaningful names.

  5. After making the clean-up, we'll see that some columns are deleted and some are renamed.

  6. When we open the properties of the logical table source named Fact_Orders, we'll find mappings from the Logical Column to the Physical column. Also if you need to access other details of the logical table source, you'll have to extend the logical table sources folder, so you'll find the logical table source definition.

  7. The new business model is almost ready but still there is a missing step that is the aggregation rule for the columns in the fact table. By default, it's not set. The following are the options for the aggregation rule, so depending on the requirement you can use one of them:

    • Sum and Avg

    • Count and Count Distinct

    • Max and Min

    • StdDev and so on

  8. In order to set the aggregation rule for one logical column in the BMM layer, you'll have to open the properties window of that column and go to the Aggregation tab. You'll find the rules in the drop-down list and select one of them. The most common one will be Sum. So we'll use Sum as the aggregation rule in this demonstration. In the case of business requirements, we can use different aggregation rules. If you need to display the number of the rows, we can use Count as the aggregation rule.

  9. So we've configured all of the measures in the fact table with the Sum aggregation rule.

How it works...

We created a simple business model in this section. Each logical table should have at least one logical table source. We might create multiple logical table sources as well, just in order to use summary tables or horizontal/vertical data federation. The important point in this layer is creating the logical joins. The Oracle BI Server will recognize the logical fact and dimension tables by checking the logical joins. Obviously, the logical joins are dependent on the physical joins, but we may have only one table in the Physical layer. So there won't be any physical join. Even in a single table scenario, we'll need to create fact and dimension tables in the BMM layer and create logical joins between them. At the end, all the objects will be referencing to only one single table. After working on the Business Model Diagram, we selected the aggregation rules that will be applied on the measure columns such as Sum, Count, and Avg. Is this a required step? Actually, it's not required. If you don't set the aggregation rule of the measures, you won't be able to see aggregated data in the reports. Let's assume that if the fact table contains one million rows, you'll see one million rows in the report. Although setting the aggregation rule is not a must, it's highly recommended to do it.

There's more...

Regarding the business model, there is one more object type called logical dimensions. The Drill Down feature and aggregate (summary) tables can be used after dimensions are created. As you see, two important features are dependent on dimensions. We're going to focus on logical dimensions and their types in Chapter 2, Working with Logical Dimensions.