compression of cube data in a process chain

Hai all,

I have a small question. If collapse or compression of cube data in process chain is included and by mistake if a wrong load was loaded into the cube, the data gets compressed automatically deleting based on requests is not possible anymore. This becomes problem and we can

t really do selective deletion for thousands of records.

So is it a good practise to include the collapsing in process chain or do it manually. Is there any other better practise?

Thanks.