IB - Developer Center    Forums  Hop To Forum Categories  Performance Management Framework (PMF)    Different Aggregation methods at different levels
Go
New
Search
Notify
Tools
Reply
  
-star Rating Rate It!  Login/Join 
Platinum Member
Posted
Not sure if this is going to make sense or not, but here goes. I'm curious if there would be a way to aggregate measures differently at different levels of dimensions. My scenario isn't really normal in the traditional sense of a scorecard, so this may not really apply to most other users.

I have an IT scorecard that we're using to benchmark application use -- specifically the number of users. In my scenario, I have multiple systems (that are set up as values in dimension 1). At the top level of the measure (no dimension visible yet), I need additive aggregation. In other words, I want to see the total number of users for all systems. But THEN when I drill into the measures and start looking at dimension 1 (by system), I need averages (ratios). Because at this level, I don't want to see all of the months add up for a single system and make it look like I have 3x as many users as I really do... Basically I don't want to double-count users across months -- but I still want to see the total # of users across systems (not the avg number of users per system).



Production: 7.6.4 WF Server  <=>  7.6.4 WF Client  <=>  7.6.4 Dev Studio
Testing: <none>
Using MRE & BID.  Connected to MS SQL Server 2000
 
Posts: 218 | Location: Wichita, KS | Registered: May 27, 2005Reply With QuoteEdit or Delete MessageReport This Post
Virtuoso
Posted Hide Post
In other words, you want to count the number of users on all systems regardless if they are duplicate users, then start drilling down into each system in use. You might want to investigate drill down reports. Have you mocked up how your report structure should appear. There are many ways to ensure you don't have a multiply effect. Are you pulling all data in the various systems and merging them or just pulling data from each that you need and then merging the information for your reports.


Leah
 
Posts: 1304 | Location: Council Bluffs, IA | Registered: May 24, 2004Reply With QuoteEdit or Delete MessageReport This Post
Platinum Member
Posted Hide Post
PMF maintains a single aggregation path along all dimensions because it's designed for that - things have to add up all the way up and down.

To do what you want would require two measures. The two different methods of aggregation couldn't be combined along a single measure.

Leah hints at the answer: design a WebFOCUS report to do this particular, different sort of thing; a WebFOCUS report drilldown can run a totally different aggregation. PMF is all made of WebFOCUS stuff too, but with a particular design to it.

In your custom report, you could pass PMF parms to it (see the latest PMF New Features doc for advice on how to do this), and make it look like the rest of the PMF reports, but this one animal is from a different breed and so needs specialized development.

thanks!


Bob Jude Ferrante
Technical Director
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105
 
Posts: 202 | Registered: March 26, 2003Reply With QuoteEdit or Delete MessageReport This Post
Platinum Member
Posted Hide Post
I suspected that would be the case. I'm definitely comfortable creating WF reports and drill-downs, so I'll keep that in mind.

We were just hoping for the ability to keep it withing the framework -- but I sort of knew the answer but thought I'd ask anyway. Smiler

Thanks!



Production: 7.6.4 WF Server  <=>  7.6.4 WF Client  <=>  7.6.4 Dev Studio
Testing: <none>
Using MRE & BID.  Connected to MS SQL Server 2000
 
Posts: 218 | Location: Wichita, KS | Registered: May 27, 2005Reply With QuoteEdit or Delete MessageReport This Post
 Previous Topic | Next Topic powered by eve community  
 

IB - Developer Center    Forums  Hop To Forum Categories  Performance Management Framework (PMF)    Different Aggregation methods at different levels

Copyright © 1996-2008 Information Builders, leaders in enterprise business intelligence.