Quando omni flunkus, mortati. Die dulci fruere.

| Subscribe via RSS

Tuesday, April 7, 2009

Go! Dashboard Install Fun


In a surprise twist I was successful at installing IBM Cognos Go! Dashboard. Why the surprise? Go! Dashboard is not one of the easiest installs from Cognos. Initially, here were the steps I followed:
  • Install Cognos Go! Dashboard
  • Change the portnumber from 8080 to 8085 in the system.xml
  • Generated the WAR file
  • Deployed it to the Tomcat server
  • Changed the Gateway URI in Content Administration
  • Generate a service

And Go! Dashboard replied: "The IBM Cognos 8 Go! Dashboard service terminated with service-specific error 0 (0x0)."

With some persistent investigation I found the following workaround:

First, I added the JAVA_HOME environment variable in the ..\cognos\c8\dashboard\bin\services.bat:

set JAVA_HOME=C:/Program Files/cognos/c8/bin/jre/1.5.0

Next, instead of generating the service, start the startup.bat file in the C:\Program Files\cognos\c8\dashboard\bin directory (default file location). After this the Tomcat server windows are spawned and Go! Dashboard actually seems to work (and everyone is happy).

Wednesday, April 1, 2009

Calculated Members


The member function is used to create a new member in a specified hierarchy. The initial argument of the member function is the expression used to create the member. The next two arguments are the business key and caption values for the new member. The final argument is the hierarchy in which your new member will reside.

A calculated member can serve almost any purpose. One example of using many calculated members is to create a Variance calculation. If you wish to calculate a variance for many different measures you wouldn't use the measures themselves. Instead, you would create a calculated member in another dimension that would form the basis of your variance calculation. Say you wanted to compare the current period values to the prior period values. This would use an expression something like:

([Current Period] - [Prior Period])/[Prior Period]

You could define this as a new member of your time hierarchy using the member function as:

member(([Current Period] - [Prior Period])/[Prior Period], 'CurPriorVar', 'Variance to Prior Period', [Cube].[Time Dimension].[Time Hierarchy])