Bulk loading into a table that already has records is a common scenario. A lot of folks ask how can they optimize this, and besides the standard ‘it depends’ answer, there’s a BOL topic that gives useful information on what to do depending on how large your new load is relative to how much data you have, and the type of indices on the table.
Sunil (who worked with my father for a brief time earlier in his career!) and Man have a TechNet article basically answering this question with a lot more rigor than I have seen in the past… the article discusses this topic in good amount of detail and provides experimental findings as well. While the doc was written in Sept. 2004, I just came across it today.
Check it out!