cancel
Showing results for 
Search instead for 
Did you mean: 

How many records in an ODS and Cube

Former Member
0 Kudos
49

Hi Everyone,

I am loading some data from R/3 to BW. The first load was a 5 day load for the month of January to ODS. The data has transformed from 2 million records to 32 million records because of a start routine. So if I continue to do loads for one full year you just can imagine what the final output will be.So want your suggestions on what are limitations for loading data in ODS and Cube and also what I should say to the customer so that he understands the loads will have on the production system. If anyone has come across such situation,please share your experiences with me.Also can someone send me something like a checklist which I can follow while dataloading(eg.TCodes, database,timing,etc)

Regards,

Kal

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

Well, it looks like it'll certainly be a large ODS 6 million+/day. If your DB server is sized appropriately, you'll be able nadle teh volume, although if you need to run lots of queries against his ODS performance could be a challenge.

On the database side (depending on the DB involved)you need to consider partitioning and parallel processing wherever possible if you are not already doing so. Is the ODS going to be used for BEx queries or just to feed other ODS/cubes?

You can also look at creating mulitple ODS, e.g. one for Q1, another for Q2, Q3, Q4, and then pull them together in a MultiProvider. You don't indicate how long any data must be retained.

You should look at creating and pushing data on to other more summarized ODS or cubes so that you can minimize the need to have to refer to this ODS.

If this a new customer, maybe they don't really grasp the implications. Guess the question is do they need to be creating this volume? You need to work with the customer to determine if the business needs can be satisfied with anything less than what you have. Is this custom data model or are you using standard business content?

Don't know that there is a simple solution to your problem at this point. Good luck.

Former Member
0 Kudos

We are not using the ODS for Bex queries so unchecked the bex reporting option.The data is for 2005.The data is being fed to the cube where these 32 million records transform to 5 million. They are using standard extractors but custom ODS and Cube.I am worried if I load this amunt of data that the system may break.