Ah hah.. well in the case of a missing PK you could make a surrogate key representing the import job / filename, etc... Your solution is ovbiously working, and performs. Like I said in the beginning, I am curious to know how my approach would perform since I believed in the beginning that it would use less memory from the lookup, but I realize now after the pivot that my solution would use more memory per column.. I'll have to do some reading about the SSIS memory "back-pressure" controls I heard about at an SQL meeting.. I think that would help keep the memory needed for the entire task to a minimum.