cancel
Showing results for 
Search instead for 
Did you mean: 

Importing Continuous New Dimension Members

Eyal_Feiler
Participant
0 Kudos

Hi Experts,

We have a requirement for a consolidation Appset to import the source system Accounts (“GL”) so they have the detail for the underlying detail.  In addition they want all the invoice detail.

We all know that this is not the purpose of BPC or any consolidation system and this is not done.  Unconvinced we need to pursue.

As a solution, I proposed that we have 2 cubes in BPC a DETAIL cube with GL and BPC consolidation dimensions and another cube CONSOLIDATION just containing the consolidation dimensions (no GL or other..).  Using BPC DM, data would migrate from the DETAIL to the CONSOLIDATION with just the relevant CONSOLIDATION dimensions

The question is:  If we want to import data   to the DETAIL e.g. GL invoice # Debit /credit etc. the dimension members need to be in place before the import.  This is not possible.

Any ideas to addressing this issue with BPC or SQL Server?


We are using BPC 10.1 MS and a toolkit (SQL Server 2012).

Thanks in advance.

Eyal

Accepted Solutions (1)

Accepted Solutions (1)

former_member186498
Active Contributor
0 Kudos

Hi Eyal,

it's not clear why you can't align master data before import, could you please explain it better?

How many dimension you need for DETAIL model? How many records you will have in this model?

Regards

     Roberto

Eyal_Feiler
Participant
0 Kudos

Hi Roberto,

The import will consist of thousands of new rows of data - one of the dimensions will be Invoice_ID.  This means that prior to the load we would need to load hundreds or thousands of new dimension members - e.g. IN_1234 IN_3456, XYZ - random invoice values - this is not practical.

Re the record count - I'm not sure but I estimate thousands at this point.

Re  dimensions - 22 dimension -12 that are true CONSOLIDATION required and 10 extra including the unique values as above.

Any ideas?
Thanks

Eyal

former_member186498
Active Contributor
0 Kudos

Hi Eyal,

if you need Invoice as dimension you have to mantain master data before import.

  • A possibility is to create a package that update and process this dimension automatically before import
  • or if you need only to read and show this data, you can also think to use an external custom table instead of a cube, buildings right indexes will help to read quickly invoices rows, and you can use this table connected with drill through functionality.

  • another possibility is to memorize invoice_id as comments.

Regards

     Roberto

Former Member
0 Kudos

Hi Eyal

I am not sure what to suggest as the best option because I don't have enough information on your requirements.

But I would strongly recommend to avoid implementing it like you described it above.

Having hundreds of thousand of members under the Invoice_ID dimension will cause trouble at many levels down the road.

First of all, it will be an administrative nightmare to update the dimension members and process the dimension on a regular basis. Then you will most likely run into every single limitation of SSAS when building the queries. On top of that you can also expect serious performance issues induced by this large dimension.

Stefan

Answers (0)