Skip to Content

Is there an Easy way to collapse with zero elimination within BPC

To make a intersection in SAP BPC as ZERO , I have to pass the negative values in the planning buffer . However these zero value records are only eliminatiion from the INFOCUBE when the cube is compressed / collapsed with zero elimination . Thus any operation via AMDP SQL perhaps reads these records and perhaps leads operations on greater data volume .

One solution could be using SUM(key figure) and grouping y the dimensions of the aggregation level with the clause HAVING Key_figure > 0

What is the solution

and is there a easier way to compless data with zeroo elimination . so that I can include this within my planning function / sequence .

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

4 Answers

  • Best Answer
    Apr 08, 2017 at 08:20 AM

    With BPC standard the approach is the same - if you clear a lot of data you get a lot of zero's. And it's a good practice to run Lite Optimize with zero elimination over night.

    Add comment
    10|10000 characters needed characters exceeded

    • Former Member

      Dear Vadim,

      Would like to ask for BPC Embedded is there a way to do Lite Optimize as in BPC Standard?

      Thanks

      regards,

      arie

  • Apr 07, 2017 at 04:59 PM

    "However these zero value records are only eliminatiion from the INFOCUBE when the cube is compressed / collapsed with zero elimination ." - and what is the issue with this approach?

    Add comment
    10|10000 characters needed characters exceeded

  • Apr 07, 2017 at 11:09 PM

    Vadim: As you correctly pointed out , there is absolutely no issue .

    However consider this .

    we have 20 K records in the infocube , and I want to make the key figures zero . I would have to pass negative of the key figure and thus the cummulative property of the infocube would automatically make them zero. But suppose another operation / planning function is called next where I have to multiply the key figure by a factor of 2 . ( i.e select dim1, dim2 , 2*keyfigure )

    So my question is

    Would the multiplying factor act upon 40 K records ( 20 K original records , and 20 K offsetting records) . which would be a waste of effort as they would end up adding to zero anyway . So a better option would be :

    select dim1,dim2, SUM(KEYFIGURE) *2 
    group by dim1 ,dim2
    having sum(KEYFIGURE) > 0  

    Compress/collapse with Zero elimination automatically cleans up my data and also free's up the RAM (how ever small it might be ). so was also wondering if I would call the zero elimination more easily / programatically

    Add comment
    10|10000 characters needed characters exceeded

  • Apr 10, 2017 at 04:41 PM

    Hi Rajarshi,

    it seems that you are talking about BPC Embedded, and thus the 'planning engine' is based on BW-IP/PAK. In other words you can also use a planning enabled DSO in the case you don't like the design of InfoCubes based on delta records; planning enabled DSO use 'after-image' logic and physical deletion is possible.

    On the other hand, talking about PAK and using HANA to compute the SUM of 20 or 40 K records is simply nothing: HANA is really good for such kind of computations. You mention AMDPs, do you want to implement a planning function based on AMDPs? If yes, there exist very good how to papers about this topic, cf.

    https://blogs.sap.com/?p=119219

    In most cases PAK does the data read for you in an AMDP, check the above mentioned how to papers and the BW-IP/PAK documentation.

    Regards,

    Gregor

    Add comment
    10|10000 characters needed characters exceeded