Hi. We run 2012 enterprise. I've never used fuzzy lookups or groupings before. All of the data is US based.
I'm faced on every row coming into my etl with the following challenge...1) is last name A equal to or similar enough to last name B to mark them as possibly the same on this row, 2) pretty much the same with A's and B's addresses. By possibly I am suggesting that perhaps I'd be able to tweak how strict the rules are based on the results I see.
I don't really know at this point if both addresses go thru some sort of US mail address standardization software but have to assume for now that the answer is no.
I read some older documentation on the fuzzy components and thought immediately that this isn't really a set type of comparison (an input set against some other input set/index) as much as it is a case by case decision with one column being compared against another.
Can the community suggest a direction. Perhaps a white paper that addresses more closely my particular challenge?
My gut says a soundex variation could be a good approach to the last name part of the challenge.