SAC Planning Model compression (import connexion)

[former_member767054]

Hello experts.

I'm meeting performance issues with an Odata export to BW, and I was asking myself how many records there is in my source model. I didn't find a better solution than exporting the entire content of my model into files.

(we are dealing with a planning model --> records are really physically stored in the cloud)

Is there a place somewhere in SAC where I could find the number of records ?

The result is around 5 Millions records (split in 5 files), but more than 80% of them are useless "0".

So here is my question : how could I drop those around 4 millions of useless "0" records ?

The Odata export log show 500 / 600 k transfered record to BW (so obviously "0"s are ignored, but maybe they have a cost in term of performance).

Thanks in advance for your feedbacks.

Rgds.

Guillaume P.

[Mastan]

Hi Guillaume,

Regarding your Question1: Is there a place somewhere in SAC where I could find the number of records?

• You can find the number of records in the source model by opening model structure workspace

Coming to Question2: how could I drop those around 4 millions of useless "0" records?

• If you know which records to delete, then you can directly delete the Facts by selecting the specific data (ex: if you want to delete the region data then select the dimension related to that and choose the region member)

• You can also use the Data action feature "Cross Model Copy" and copy the data whichever is useful for you into the other model of same structure.

Thanks,

Mastan

[former_member767054]

Hi reddy07 .

Thank you for the answer, this is useful.

By chance I don't have too much different versions so far (not visible in your screenshot but on the right you have to restrict the versions dimension, then you sum up together number of records per versions).

About the 2nd point : unfortunately there is no way to locate "0" using dimension filters. They are spread more or less everywhere.

Rgds.

G.

[marcelo_portella]

I would say the best approach is to drop the 0 amount records before recording them on the model.

If you are importing data from a source system, it is possible to do it during the data preparation.

Just find a record with 0 amount, click on the cell and select the "Delete all rows with 0" option.

Cheers!