Just a thought. Wouldn't it be easier to simply pull the entire file into a single staging table, and then manipulate it out of that table into whichever destination tables you require. Presumably this would be easier than going back and forward to the flat file for every single column?