Skip to Content

Hello SCN,

In this blog, we will discuss about the structural changes of In-Memory Cube when compared with Standard Info Cube and also discuss about different advantages and limitations of using In-Memory technology with Cube. This blog is in continuation with my previous blog: *BW on HANA: Working of In-Memory DSO (BW on HANA: Working of In-Memory DSO)

Let us now discuss about existing schema of Info cube:

Extended Star Schema of Standard Info Cube:


For detailed understanding of extended star schema, please read my wiki: Extended Star Schema

How Does In Memory Cube look like?

As we have an additional boost of HANA computation power. Dimension tables are removed and Master Data tables are directly linked with F table.

There is no more E table. You can see the structure of In-Memory Cube below:

How does a Fact table of In Memory look like?

The fact table (F) no longer contains Dim Id (Except packet dim id). We can now see SID’s in Fact table as shown below:

Dictionary of Standard Info Cube Tables:

Example: We can see in SE11:Or we can use the Report “RSDRI_CONVERT_CUBE_TO_INMEMORY”.
What does this conversion do?

Thank you for reading this blog πŸ™‚

To report this post you need to login first.

20 Comments

You must be Logged on to comment or reply to a post.

  1. Mark FΓΆrster
    Hello Krishna,

    thanks for the blog. I can’t wait seeing a BW 7.3 on HANA myself. And somehow I really need to see it, because I can’t believe that SAP finally simplifies something. For the first time in known history! (as far as I know)

    If data loading is significantly faster than with traditional RDBMS/InfoCubes then that would amount to a massive advantage even for BW customers who already have a BWA.

    Cheers,

    Mark

    (0) 
    1. Lalitha Swaroop Krishna Tangudu Post author
      hello Mark,

      Thanks for your feedback. HANA is a really cool. As a HANA tool, it helps you in faster loading and reporting as well and is also helping many projects to perform complex analysis which earlier took a long time.
      And with respect to BW,Use case of Redbull already proved its loading performance.

      regards,
      Krishna Tangudu

      (0) 
  2. Zahir Malpekar

    Hi,

    In-memory optimized infocube has many advantages

    1. No longer dimensions required. Masterdata is directly linked to fact tables. This helps to really faster the process of activation of infocube upto 5times faster data loads.

    2. Simplied data modeling

    3. Faster remodeling of structural changes i.e when key figure is dropped, the infocube is to be unload, make structural changes, reload the infocube & activate but using in-memory optimized infocube, just dropped the keyfigure & you are ready to go without any activation. Even you can add the new keyfigure without unloading the data but of course you need to populate the new field.

    4. No changes required in exisiting proces in BW. Multiprovider, Queries remains unchanged against the infocube which is to convert.

    5. It reduce the maintenance & total cost of ownership

    6. No replication i.e. faster query access directly on primary data persistence

    7. Indexes on Infocubes, Infoobjects no longer required (no Rolls -up, change runs)

    8. Many calculation is pushed down to in-memory such as currency conversation which cannot be pushed to database using BWA.

    Regards,

    Zahir Malpekar

    (0) 
    1. Lalitha Swaroop Krishna Tangudu Post author

      Hello George,

      Thanks for reading the blog. As far as i know, there is no limit on Max number of SIDs to be used. But i would request other members in the community to throw some light on this if we do have any limitation regarding the number of SIDs.

      Regards,

      Krishna Tangudu

      (0) 
      1. JITU P

        HI Krishna,

        the data in the key field of the fact table is repeating and more ever it contains unique values equal to the number of charcterstsic.

        could you please explain how the data stores wigth a examaple and how the key filed value will be populated while loading data to the cube.

        Thanks,

        Ranjan

        (0) 
  3. Sreenivasulu Reddy

    Hi Krishna,

    I read cube model in HANA environment.It is very nice.But I would like to know

    1).The primary & Foriegn key relation ship between fact_sid and Masterdata.(With example)

    As you are saying sid is in Fact. When will loaded SID table?

    2) In extended star shema we can include master data tables as much as we want with help of Dim Id’s.How can we achieve same in HANA?

    Thanxs

    Sreeni.

    (0) 
  4. Viren P Devi

    Hi,

    Quick question ,

    Now as we have SID id’s directly in Fact table , do we not have limitation of using only 16 key fields.i.e 16 chars. I do not think SAP would let that happen. What am I missing here?

    Regards,

    Viren

    (0) 

Leave a Reply