Guideline

Setting Up Dimensional
Information in Framework Manager
Product(s): IBM Cognos ReportNet 1.1
Area of Interest: Modeling
Setting Up Dimensional Information in Framework Manager 2

Copyright
Copyright © 2008 Cognos ULC (formerly Cognos Incorporated). Cognos ULC is
an IBM Company. While every attempt has been made to ensure that the
information in this document is accurate and complete, some typographical
errors or technical inaccuracies may exist. Cognos does not accept responsibility
for any kind of loss resulting from the use of information contained in this
document. This document shows the publication date. The information contained
in this document is subject to change without notice. Any improvements or
changes to the information contained in this document will be documented in
subsequent editions. This document contains proprietary information of Cognos.
All rights are reserved. No part of this document may be copied, photocopied,
reproduced, stored in a retrieval system, transmitted in any form or by any
means, or translated into another language without the prior written consent of
Cognos. Cognos and the Cognos logo are trademarks of Cognos ULC (formerly
Cognos Incorporated) in the United States and/or other countries. IBM and the
IBM logo are trademarks of International Business Machines Corporation in the
United States, or other countries, or both. All other names are trademarks or
registered trademarks of their respective companies. Information about Cognos
products can be found at www.cognos.com
This document is maintained by the Best Practices, Product and Technology
team. You can send comments, suggestions, and additions to
cscogpp@ca.ibm.com .

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 3

Contents
1 DIMENSIONS ................................................................................................ 4
1.1 UNDERSTAND THE DIMENSIONALITY OF YOUR DATA. .......................................................... 4
1.2 IDENTIFY THE LEVELS, LEVEL KEYS, LEVEL ATTRIBUTES AND HIERARCHY(IES) OF YOUR DIMENSION. 4
2 DIMENSIONAL INFORMATION AND SQL GENERATION ............................... 5
3 OVERVIEW OF DIMENSIONAL INFORMATION COMPONENTS AND THEIR
IMPACT ON REPORTNET .......................................................................................... 6
4 SPECIFY DIMENSIONAL INFORMATION IN FRAMEWORK MANAGER.......... 7
4.1 PROVEN PRACTICES TO FOLLOW ................................................................................... 8

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 4

1 Dimensions
Dimension are used in the following way for ReportNet.
1. Associate measures with levels to ensure proper aggregation
a. For example, to avoid double counting in queries
2. Define keys for groups for query efficiency
3. Model aggregate tables
4. Model multi-grain aggregate tables
Note that dimensions do not appear in the Report Studio and Query Studio user
interfaces as they are intended for query planning and not for user exposure.

Dimensions are optionally defined on each query subject and use only the query
items defined therein.

1.1 Understand the dimensionality of your data.

When specifying dimensional information in FM it is important to understand the
following.
• What are the levels in your dimension?
• What is the order and combination of levels that form hierarchies.
• How do you uniquely identify the members of a level, is it unique (one
query item) or does it require multiple query items?
• Association of levels with attributes and measures.
• At what level of granularity is your fact data, do you have more than one
level of granularity? Ie. Is some of your data recorded monthly and some
daily, this would mean that you would have a fact tables relating to your
time dimension table at the month and day levels.
• Are there surrogate keys defined in your data source?

1.2 Identify the levels, level keys, level attributes and hierarchy(ies)
of your dimension.

A star schema might have a table that looks like:

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 5

It can be observed from this data that product lines have multiple product types
and product types have one or more products. The cardinality of the product
line-product type relationship and the product type-product relationship should
be verified. A hierarchy that might be useful for reporting would have the levels
Product Line, Product Type and Products and would be structured so that
Products relate to and are children of Product Types, and Product Types relate to
and are children of Product Lines. Once it is clear what the levels are and what
the relationships are between the levels it is important to determine what
uniquely identifies a level (the key) and which data elements are associated to
each level. In this case Product_Line is a unique identifier for the Product Line
level, Product_Type is a unique identifier for the Product Type level and
Product_Key is a unique identifier for the Product level with Product_Number and
Product_Image as attributes. You may wonder why Product Key was chosen as
the identifier for the Product level instead of Product Number. The answer is
that the datasource is a warehouse that uses surrogate keys and Product Key is
the surrogate key for the Product dimension table. Product Number is also an
identifier but it is regarded as a business key and will be treated as an attribute
when specifying the dimensional information in Framework Manager. The
dimensional information is summarized in the following table:

Levels Product Line Product Type Product Name

Keys Product Line Product Type Product Key
Unique yes yes yes
Attributes N/A N/A Product Number,
Product Image
Hierarchy Product Line Product Type Product Name

2 Dimensional Information and SQL Generation

It is important when specifying your dimensional information to understand the
affect this will have on the SQL that is generated. Dimensional information
affects grouping and aggregation of data, not only from the dimension data
source but also from other data sources that have relationships with that
dimension. Specifying hierarchies for a query subject using query items as
levels, either individual or combinations of query items as keys, and associating
query items to the appropriate levels as attributes provides FM with enough
information to automatically group and aggregate data according to the
hierarchies you have specified. Hierarchies in this case can be thought of as
ways of specifying the order in which data segmented into levels can be
grouped.

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 6

With the dimensional information specified in the model it will no longer be

necessary for a report writer to understand the relationships between the data
elements and do explicit grouping in a report in order to get correctly aggregated
data. The model will have sufficient information to write the correct SQL so that
data joined on Product_Key can be aggregated properly when reported by
Product Line or Product Type or both.

3 Overview of Dimensional Information Components and

their impact on ReportNet
The key is optional and used to identify a query item(s) that guarantee
uniqueness. By default the query item that is the level is set as the key. A single
query item key is used when one item uniquely identifies the level - for example,
product number uniquely identifies each product. . If a single query item
determines uniqueness, then use the Unique Key option to inform ReportNet.

Multiple part keys are used when two or more query items are required to
uniquely identify the level - for example, the combination of country and city are
required to uniquely identify each city. Note that each level must be uniquely
identified on its own, not by other levels within a hierarchy.

Facts are associated with levels to tell ReportNet which quantitative query items
can be aggregated against that level. For example, if a query subject had the
country and city query items and a population column, ReportNet would need to
know whether the population was for the country or the city. Where facts are
stored at different grains (for different levels) in the same table, ReportNet
needs to know which facts are associated with each level. For example, a query
subject with country and city might have two population query items one for
country and one for city. By properly associating the fact with the appropriate
level in Framework Manager, you ensure that ReportNet will correctly aggregate
the facts avoiding such issues as double counting.

Attributes are qualitative columns which share a 1 to 1 relationship with the

level. For example, a product description exists for each product. Attributes are
columns which are generally not aggregated.

Once levels have been defined they can optionally be used in hierarchies.
ReportNet will attempt to derive hierarchies if they are not set. The hierarchies
establish the relationship between the levels - for example, a hierarchy of
country - city tells ReportNet that country is a higher level grouping to city.

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 7

4 Specify Dimensional Information in Framework Manager

Before Specifying Dimensional Information in FM ensure that you have verified
your relationships, query item usages and aggregation properties.

In FM, dimensional information described above would look like this.

IBM Cognos Proprietary Information

Setting Up Dimensional Information in Framework Manager 8

4.1 Proven Practices to follow

Below are some proven practices to follow when Specifying Dimensional
Information in Framework Manager

1. Ensure that keys and other identifiers have their usage specified as identifier.
2. When a surrogate key is available for a level use it as the level key and treat
the business key as an attribute.
3. Use the item you would use in reporting as the level and the id (forms one
end of a relationship) as the key for the level (remember to reflect this in the
prompt info for the sake of query efficiency).
4. For each level associate any query items whose granularity is specific to that
level.

IBM Cognos Proprietary Information