I agree with Phil's approach.

SELECT

EXCEPT

SELECT

is a tried and true way of doing a column specific "table diff". I use hashing as a last resort for extremely large and wide dimensions with many type-2 columns. Personally, I love Todd McDermid's SCD Transform (free on codeplex), but in this situation, simplicity rules...especially if there are not many on your team that know and can support SSIS.

Unfortunately, SQL Server does NOT include the ANSI standard "CORRESPONDING BY" functionality yet....maybe someday....