If you are unable to create a new account, please email support@bspsoftware.com

 

News:

MetaManager - Administrative Tools for IBM Cognos
Pricing starting at $2,100
Download Now    Learn More

Main Menu

most efficient way to build a cube report

Started by kiran.timsina, 25 Feb 2013 09:44:21 PM

Previous topic - Next topic

kiran.timsina

My client is complaining about the performance of cube report. It takes 4 or 5 minutes to drill down to subclass level from class level, which is quite disgusting.
The development tools I am familiar with are framework manager, Report Studio, Analysis Studio.

Based on my research and the tools I am familiar with, I am considering to
1. use IBM TM1 as OLAP data source (is TM1 better than DQM data source?)
2. Define a TM1 data source (earlier I had been using CQM, DQM enabled data source)
3. Model a DMR package in framework manager
4. Use Analysis Studio for cube report or even Report Studio for some drill down reports

Star schema(with the dimensions snow-flaked) data warehouse in on Netezza. Suggestions outside the tools I'm familiar with will also be appreciated.

RKMI

Hi,

I don't get the point of this topic, if you report drill downs are inefficent why would you change you technology...

Have you explored the query/ ran a execution plan? How the joins between your fact and dimension tables are made if they are Db cost effective? Are you doing alot of rollup calculation in which case it might be better to introduce a summary table? Have made sure you have the necessary keys to optimize the path? What is the size of fact table? I mean there are so many unanswered questions before you go down the path of exploring a new OLAP technology...

Thanks,
RK

Rahul Ganguli

I agree with RKMI.
So you have DMRs or Power Cubes?
If it is DMR, you have to optimize the underlying relational database, but since you are using Netezza I think the query generated may have some issue.
Regarding your first point "is TM1 better than DQM data source?", DQM is not a data source it is an access mode for the datasources, DQM mode is available with all the datasources in COgnos you just need to enable them. But, maximum gain is achieved in multi-dimensionall data sources.
On your third point, IF you are using TM1 as data source you need not to model in Framework Manager, you just need to publish the package.

Hope this helps.

Regards,
Rahul