group by + having

is your friend.

E.g.

Select field1, field2, count(*)

From tablex

Group by field1, field2

Having count(*)>1

But in your scenario i do not know if this is the solution. Question is if record was changed and in next update "restored" and once again changed is this duplicate or no?

If yes above query is ok if no then procedure is your friend.

And for the future change your audit trigger and check before if there was any change

Thank you, but I this isn't quite what I need.  I need to match against all the records of a set (up to 8 rows per set) - not just individual rows.

--

Daniel