Lag function in TM1

[newtotm1]

Cognos Planning has a BIF called Lag this will allow one to lag the data within a cube by one month or x months.  Is there an easy way to do this in TM1?  I know I can do this by creating an attribute and then defining the what the lag is but this is very cumbersome way of doing because time dimension is 60 months.

[OLAPBPMguy]

Sorry to sound trite but if you know how to write rules this should not be cumbersome.  Rules to lag or spread input data over many subsequent time periods are very common in TM1 models for example asset depreciation, recognition of subscription income (pay TV, magazines), insurance underwriting, ... etc.

The number of time periods is not really relevant.  Depending on the problem lookup cubes could or probably would be better than attributes, but it will depend. Sometimes an auxiliary calculation cube can also be helpful, but in the main this type of calculation is straightforward and not cumbersome to implement.

The most important thing is you need to start thinking about the business problems and calculations MULTI-DIMENSIONALLY as opposed to DIMENSIONALLY as you are used to doing in planning.

[dusherwo]

I agree it is doable (hell, I do it all the time  :) ) but I do think it could be a lot more straightforward than it is.  TM1 doesn't really 'do' time series intelligence at all.
A related issue is that (mostly) BI and EP designs have a single time dimension. TM1 designs can use a single time dimension but often have two time dimensions (Year and Month) (or even more  :) ). Using two time dimensions makes YTD very easy and leads/lags rather more fiddly, especially if you want to look back/forward more than one period.
At present, in modelling, I use a mix of dimensionality, holding input data by year and month, using a flat period dimension for the forecast and summarising the output back to year and month. Took some setup but has proved quite portable from client to client.

[AWILDE]

Hi, you can do this by either using an attribute or looking at the index number on the dimension - the rule is straight forward and simple to write.  By not having a set time period in TM1 this allows opportunity to make your time periods work the way you want them to - e.g. multiple dimensions to control your time etc which can improve reporting etc.