The fastest way is probably to not load the excluded rows in the first place:
[Exclusion Table]: LOAD Field1 & ',' & Field2 as ExclusionKey ... ; [Main Table] LOAD UniqueID ,some fields ,Field1 ,Field2 ,some other fields ... WHERE NOT EXISTS (ExclusionKey,Field1 & ',' & Field2) ;
If you MUST load the main table first (or if using QVDs, where the not exists above would, I believe, cause an unoptimized load), probably something like this: