I try to find a way to analyze table uniqueness that doesn't have primary key defined. I try to find an approach without db server resource more like query/read only to be able to run uniqueness queries to find out the column or composite column creating a unique record.
Thought of running several distinct count vs. count on each column but it will be very heavy and will not solve combination of keys. might use log2 search but it seems also heavy
Any ideas, SQL code to support that need?
Thanks!