TM1 Forum

Hi

Using 8.4.2 and Excel 2003

I have a dimension with 1.5m elements. No hierarchy other than a total items - its a list of jobs raised over the last 5 years

The element attrubute cube is 161MB

The cube using the dimension is very clunky when using the cube viewer. Its 68MB

Any thoughts on easing the burden ...on the face of it we have 4Mb of RAM and half of it is in use (IT won't let me see the task manager to view any spikes)
I've optimised teh dimension order, so think I've gone through the right processes

thanks for your help

David

I'd be looking at an archiving strategy. Sounds like somewhere along the way you are storring transactional or quasi transactional data.

One approach (of many) would be ...
If individual job numbers are no longer needed for slice and dice capability for historical data past a certain point then collapse the old job numbers into time period archives, export data, reload to the time period dummy then delete the elements. This will shrink the dimension enormously and give much better performance accordingly. Note that with other dimensionality in the cube I think it's probably pretty likely that you could have a simple string cube to hold the actual job number(s) for each data point. You coudl then use this information to drive a transactional drill-through in case you are concerned with losing any of this type of functionality (especially in EV).