Updating historical data from single characteristic

[Former Member]

Hi mates,

This is the scenario in my client:

We have a cube with 0SOLD_TO. This characteristic wasn't linked in the transformation from datasource since the go live. Instead, it has been linked a year or so later. Now, we have a lot historical records with 0SOLD_TO empty, but we do have the registers on the PSA.

Which would be the best way to populate the records with the empty 0sold_to?

- We tried to upload the historical records with 0SOLD_TO filled into a DSO. Then, we made a self-loop transformation to the infocube, seting the ratios to 0 and making a lookup to the DSO in order to get the 0SOLD_TO. Then we tried to compress the cube but the records won't compress. This just doesn't work.

- We considered the remodeling tool but it does not apply since there is not a char or a keyfigure to add, modify or delete, it's just its historical data we need to fill.

Any advice would be very useful.

Thank you very much in advance,

Cheers.

[sander_vanwilligen]

Hi Ariel,

I suggest to delete and reload data. However, it is not clear for me what kind of data flow it concerns.

If you could provide more information on the data flow, e.g. DataSource, DSO and InfoCube, then I am able to make it more specific to your situation.

Best regards,

Sander

[Former Member]

Hi,

Create a self-loop on the cube to populate the sold_to making a lookup on the DSO and negate all the key figures of the exsiting record. Means your code should add new records with sold_to and negate key figure values for all the incoming record. Then do compression with 0 elimination. That will solve your problem. You can perform a selective full load or full load.

Regards,

Aravind.