Hi Madhu,
I am only familiar with the option "unique data records" in write-optimized DSOs. In this case the system will generate a unique index (based on the semantic key) to make sure that no duplicate records are loaded.
In a standard DSO the records are unique by definition since the key fields will be respected. This is standard system behavior, no need for any other setting.
Maybe I don't understand completely your scenario. Do you use the DSO for real-time data updates (e.g. Integrated Planning)?
Best regards,
Sander