3 Replies Latest reply: Jan 22, 2013 10:33 AM by 770149 RSS

    Question about Agg

    770149
      In one of our essbase cube, we have 10 dimensions.

      Months Dense,
      Years Dense,
      GrossNet Dense,
      Scenario Dense,
      ActYears Sparse,
      Accounts Sparse,
      ActMonths Sparse,
      Vendor Sparse,
      AFE Sparse
      Report Center Sparse,

      Some of the data is being loaded to the top level of account dimension, ie loaded to accounts. The outline of account dimension has a plus sign all the way up. Then an Agg is performed on account dimension. I would assume all data will be wiped out since there is no level 0 account member data. However the data on the top level of account dimension is still there after the Agg. Why? BTW this is on Essbase 11.1.2.2.

      Thanks
        • 1. Re: Question about Agg
          TimG
          Depends on your AGGMISSG setting, but by default #Missing will not consolidate.

          See:

          http://docs.oracle.com/cd/E26232_01/doc.11122/esb_tech_ref/set_aggmissg.html
          http://docs.oracle.com/cd/E26232_01/doc.11122/esb_dbag/dcaoptcs.html#dcaoptcs61358
          • 2. Re: Question about Agg
            CL
            Why go through all of this? If you took those upper level loads in Accounts (fwiw, I almost always make dense dimensions fully dynamic) and stuck them into level zero buckets (No_Account, No_Whatever) then all of this would be moot and the data would be there, always no matter what you did with an aggregation.

            It's a little awkward but really very easy to manage once done. I've never had a user that was confused by it.

            Regards,

            Cameron Lackpour
            • 3. Re: Question about Agg
              770149
              Thanks Tim and Cameron. I think it is the SET AGGMISSG ON issue. And I agree with Cameron on the no_account design. It may take a while for me to convince my lead to change it. Also I am puzzled on why having 4 dense dimensions. Normally account and time dimension should be dense and others should be sparse. We are having some performance issue because of 4 dense dimensions make the database very big.