How to avoid duplicate records in cube

[Sarayucog]

Hi

I have a cube and I had implemented the incremental updation option. Its running fine.

My problem is if I run the cube twise or thrise in a day its fetching all the data means its duplicating the existing data. I don't want to do like that.

For EX: My database have some 6000 records.

I am running a cube at 9:00 AM on 11-june-2010, it has fetched 3000 records means that 3000 of records entered into cube and due to some network problem its got disconneted. If I run the cube again on the same day it should fetch from 3001th record, but not from the 1st record.

So that I want to avoid the duplicate values in the cube.

How to do this could you please tell me this. Its urgent...



Many thanks in advance
Sarayu