I'm not a big fan of MERGE. I much prefer REPLACE INTO. You code everything one way. It looks just like a regular INSERT statement except for the word REPLACE. The match is done on the primary key (you don't use auto number fields as primary key, right?). Rows affected comes back 1 if it worked, 0 otherwise. Neat and clean.

If I know that I want to update I use UPDATE and check Rows Affected. If I know this is a new record I use INSERT and let the collision prevent duplicates. There are times when I have to update and/or insert a stack of data and I don't want to mess around about it. REPLACE INTO cuts my work load.