Dimension question

[job67]

I have a question regarding dimensions and, as I'm very new to TM1, I thought it might be easier to ask a question straight away rather than banging my head against a wall for 2 days and then asking for help.

We have cubes set up for different parts of our business and each is made up of the same number of dimensions. Some dimensions are generic and are exactly the same in all cubes but some are cube specific, albeit with basically the same information but just business area specific.

One of these such dimensions is cost centres where we do have an overall list in a cube for overall totals from the other cubes. However, there are then a number of dimesnions each containg a business specific subset of this overall list.

The problem is that when we have to do any maintenance with cost centres (adding new ones and removing closed ones) this necessitates changing both the overall dimension and the business cube specific dimension.

This seems to me to be a duplication of effort and risk prone if both dimensions are not updated.

So, I was wondering if it is possible to have one dimension with the overall list of cost centres which could be part of the overall totals cube as it is now. But, rather than then having separate smaller cost centre dimensions for each cube, include this overall list dimension in the business area cubes but then limit the cost centres that are shown to only be those relevant to that business area.

Hope that makes sense but I will post mre info if required.

Apologies if this is straightforward but I couldn't find anything via a search.

Thanks

Elton

P.S Forgot to say that my initial thought was to do this by creating dynamic subsets (using MDX??) by checking which business cube was being viewed and then building a subset based on the cost centres relevant to that business area. Thats's probably not possible but thought I'd mention it.

[ajain86]

Elton,

had to do something similar. I created 1 dimension with everything and share it will all cubes. Then created a different subset for each cube. The users are able to see the entire dimension if they wish, but with security you can control users from looking at data outside those subsets.

[lotsaram]

You could use element security to limit the cost centers that users have access to to only those that are relevant to them. Then you can use a single cost center dimension.

However since it seems that data is in separate cubes this may not be necessary. If you are not concerned that users can see cost centers that are not relevant (as opposed to actually seeing data attached to them) then setting appropriate subsets to make navigating the dimension easier could be all that you need.

[Steve Rowe]

A slightly different take on this since it sounds like the cubes are already built and populated. If you take the appraoch suggested then your going to have to rebuild all the cubes.

Against your master dimension have an attribute to say which cost centres relate to which business area. Then have a TI which will update all of the sub-dimensions. This means that you minimise the risk due to human error (provided you run the TI after updating the main dimension) and don't have to get into rebuilding cubes.

If you're new you're going to need to ask for help on that too but feel free to do so!

HTH
Cheers

EDIT - Grammer, spelling etc etc from posting before first coffee has kicked in

[job67]

Many thanks to all for your suggestions.

We do indeed have cubes already built and populated but we are just about to embark on redesigning our cubes, so I will probably need to look at both approaches to be honest.

As Steve alluded to, there may well be more questions to come. However, I'll spend some time trying to work some of it out for myself first though before admitting defeat and asking the forum!

Thanks again.