Noncumulative Cube Copy for Inventory IC where compression is not possible

[Former Member]

Dear All

We have got Month (0RT_C37) and Week (0RT_C36) Inventory Infocubes. Compression has never been done on any of this cube. The data is fine and result is coming correct but output is very slow. We have done the initialization on this cube on mid of 2008. We have to copy the entire cube to a ZCube and do the compression performed on that Zcube. Our system is BW7.3 SP Level11 and we are going to upgrade to 7.4 .  We tried to move the marker values(RECORDTP=1) to the Zcube with DTP setting (Initial Non-cumulative for Non-cumulative values) which is not happening. But other values with movements we are able to move. The only OSS note found to address this issue is1426533 which is not relevant for us. Is there any workaround for the situation where no compression is ever happened on the cube and the compression is impossible due to some issues, hence we want to copy the entire data to another Zcube and perform the compression there.

Regards

Reshoi

[Former Member]

Dear Sander

I tried with the solution which I was mentioning and got the result properly. I have run delta also on top of that and the result is correct. Only Issue I got is with the cardinality table correction which was solved through an OSS note. Could you please let me know how we can minimize the downtime if we are going to implement the inventory flow from the scratch also the implementation of the DSO for the movements. Currently we have modified the standard Infocube with an additional characteristic for Movement Types. Is there any standard DSO available for the movement types.

Regards

Reshoi R

[sander_vanwilligen]

Hi Reshoi,

There is no DSO for 2LIS_03_BF in Business Content as far as I know. It is possible to create a DSO yourself. Please review SAP Note 581778 - ODS capability of extractors from inventory management and the related SAP Notes carefully if you want to go into this direction.

Please have a look at the document Re-initialization of the Material Stocks/Movements cube (0IC_C03) with 2LIS_03_BF, 2LIS_03_BX and 2L... for contains a step-by-step instruction and lots of valuable information.

Re. the downtime, there are ways to minimize it. Please see SAP Note 753654 - How can downtime be reduced for setup table update. You can dramatically reduce the downtime to distinguish between closed and open periods. The closed periods can be updated retrospectively.

Please refer to document How to Handle Inventory Management Scenarios in BW (NW2004) for an evaluation of different Inventory Management scenarios.

I also would like to recommend you to review SAP Note 419490 - Non-cumulatives: poor query performance for 5 performance tips.

Last but not least, you might want to have a look at the following SAP Notes:

• SAP Note 436393 - Performance improvement for filling the setup tables;
• SAP Note 602260 - Procedure for reconstructing data for BW.

Best regards,

Sander

[Former Member]

Hi Sander

Thanks for the valuable informations and links and sorry for late reply as I was busy stuck with some urgent works. Will go through all those links and will see if it helps in our case or not. One more thing I wanted to clarify is that, Is there a way to initialize the stock based on the other cubes.  Eg. We have Daywise inventory cube (0IC_C03), which has no data and Monthwise(0RT_C37) and Weekwise(0RT_C36) cubes which has got correct data. Is it possible to initialize the cube 0IC_C03 based on the current stock of these month/week wise cubes and then continue filling the cube 0IC_C03 with delta data based on either the cubes or the DS 2LIS_03_BF

Regards

Reshoi R

[sander_vanwilligen]

Hi Reshoi,

Theoretically speaking you could simulate a 2LIS_03_BX stock initialization with marker update using BW data. However, I never have seen such an approach in practice. I advise you to be very careful with any "non-standard" approach.

Anyway, it's an interesting idea. If you would go for such a scenario, please keep us informed !!

Best regards,

Sander

[Former Member]

Dear Sander

I have successfully performed a backup using a single article to another Zcube and the data seems to be correct. I will perform the same with more articles and see if it is going right. If so then I will come to the conclusion that what I am currently doing the solution for this.

Regards

Reshoi R

[sander_vanwilligen]

Hi Reshoi,

Please keep us informed on the progress and if this will be your final solution approach.

Best regards,

Sander

[sander_vanwilligen]

Hi Reshoi,

Inventory InfoCubes must be compressed to guarantee query performance. Why did it never happen in the past? What are the issues which make compression impossible?

Please give a bit more background information to understand your situation better and why compression should not be possible.

Thanks,

Sander

[Former Member]

Dear Sander

Our BI System was initially implemented on top of an R3 system. Later it was migrated to a fully Netweaver Platform. This was done using an SAP custom tool and the migration partner was SAP. Unfortunately Inventory cube was not compressed at the time of migration and when we tried to compress after the migration it is not happening. Hence we decided to move the entire data to a Zcube and then perform compression. For that First we need to move the data with marker values (0RECORDTP =1) & compress those requests and then transfer the rest of the data with movements (0RECORDTP = 0)

Regards

Reshoi R

[sander_vanwilligen]

Hi Reshoi,

Thank you for your detailed information. I understand that you have a very special case.

Please test such an approach very carefully. I hope it will work out well.

Alternatively, did you consider to build up Inventory Management from scratch? There are ways to minimize the downtime and it would give you the opportunity to improve the data model and data flow. E.g. I recommend to add a Goods Movements DSO into the data flow.

Best regards,

Sander