on 01-08-2015 9:04 PM
Hi All,
Data load from DSO to Cube is taking 2 hours for loading 2 Lakhs records.While checking the monitoring screen, Conversion of Characteristic Values to SIDs is taking more time. Right now, SID Generation is checked in DSO. While checking the SAP_INFOCUBE_DESIGNS, one the dimension is showing more records than Fact table. Right now we are facing this issue in Production, Can you please suggest how to minimise the data loading time from DSO to Info Cube.
Thanks.
Hi Jalina,
1. Check what object do not repeat data in dimension (0doc_num, 0prodorder, objects unique wich contains attributes).
2. Create a new dimension and mark Line Item Dimension
3. Move your characteristic here
4. Transport and run a test in your QA system
5. Transport and reload your cube in PRO.
Check in google how to use sap_infocube_design, you can find many info.
Regards.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Franc,
I agree with your Points. But right, the Info cube contains 2 years data. And right now, we are unable to remodeling the Info Cube. Can you please guide me what are the other options we can do without deleting the data from the info cube. And how to improve the data loading for this info Cube. 1st is Increase the parallel processing in the DTP level. 2nd is Number Range Buffering. Apart for this, is there any other way to improve the data load.
Thanks.
Hi Jalina,
Your queries works with Multiprovider?, if yes, i think you need create archiving with new cubes for actual year and actual year - 1.
That cubes must be in Multiprovider, and this is a simple way to access your data, only need load your cubes (actual and actual-1) with your historial cube.
--> Cube Actual Year
Datasource-->ODS --> Historial Cube (Others year) ---> MP --> Querys
--> Cube Actual Year - 1
Please, share what is your flow?.
Regards,
User | Count |
---|---|
69 | |
7 | |
7 | |
6 | |
6 | |
6 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.