Remove possible duplicates based on a query.
For example, I might have a query of records with no addresses. Or I might have a list of records that I recently added relationships between and now want them re-scanned, taking the relationships into account this time.
I normally choose to scan "All Records" and include only records that have not yet been scanned or have changed since the last scan. I want these records to be marked as "scanned" but I don't want to see them in the list of possible duplicates!