Remodel

Hi Dona,

Suppose you have lots of data in your datastore coming from more difficult and timeconsuming extractors, it might be worth to consider a second option I have indicated.

For example, suppose you need to reload all sales order data for 3 years from the source system. You will first need to schedule setup jobs to collect the data in the source system and subsequently transfer this data to BW.

This collection of data and the subsequent trasnfer can be very timeconsuming.

In this case I would recommend to make a copy of your datastore, make update rules or transformation to it (this will be just 1:1 easy mapping) and load your old data into the copy DS. After succesfull data transfer you delete the data in your original DS, and remove the field. Create update rules or transformation from copy to original and load data back into the original datastore.

However, I would only do this (copy) in exceptional cases.

E.g. it could also be timeconsuming to transport to copy model to production and cleanup afterwards, depending on the rules in your company. (Of the record : if you cannot create this copy model direct in production, I find it too much of a hastle to transport...).

Also some technical impact to be considered is that you will have only 1 request ID left in your DS after the copy and that it will be more dangerous to activate an enormous dataset in one go (might need extra system resources)

Hard to say when you could consider to copy and copy back.

Using the standard is always best, unless you need to load 15 million order of course....

Hope I did not confuse you too much...

regards,

Olav