I am new to Cognos and want to know if anyone has any documentation on the structure of the Impromptu Catalogs. I am trying to understand the way they need to be built. Any examples or documentation would help.
Thanks in advance.
No specific documentation on what the catalog structure should be, but there is books about creating catalogs that you would want to read.
I Have found that putting your folders together like this makes things easier to create efficient reports. Having your Fact tables in one folder, and your Dimension tables in another folder. However, you may want another way to make it easier for users. It depends on who is using the catalog and their level on the knowledge of the data for creating impromptu reports.
FACT US
Fact US Flagpoles Monthly
Fact US Guns Monthly
Fact US Gear Monthly
FACT Canada
Fact CA Flagpoles Monthly
FACT Europe
Fact ER Flagpoles Monthly
Fact ER Gear Monthly
FACT South America
Fact ER Flagpoles Monthly
Fact ER Guns Monthly
DIM
Dim Customer
Dim Region
Dim Products
Dim Dates
I agree with CognosProfessional, it depends on who is using the catalog. We have the data separated into folders by subject area and within that folder separate the facts from the dimensions. We go another step, as we have multiple subject areas we have common dimensions in a separate folder from subject area specific dimensions. the result looks something like this,
Sales Subject Area.
Sales Facts
Sales Key Dimensions
Product Dimension
Time
Customers
Database Tables (this folder contains all tables and is hidden from normal users)
As we build these subject area folders from numerous database tables we hide the background work from the users. There is a bit of work when adding new fields etc to common folders but the payoff is a much more user aligned solution. The one thing we do stress to the users is that they stay within 1 subject area when selecting fields for their reports due to the way the tables are joined. Eventually we will have all of this controlled via Architect but at the moment it doesn't seem to handle Oracle Materialised Views very well and our 2 largest dimension tables are built this way.
It can take a while to get a catalog set up correctly and might need several iterations to get to a workable solution but is time well spent if it helps in the take-up of the system by the users.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.