in reply to DBI Question: bulk insertion with Unique-key constraint

How about LOAD DATA INFILE '...' REPLACE INTO TABLE '...'?

I also have a question about your schema... you already have a primary key (the auto_increment column.) Do you need a unique constraint on the other column set? If so, is REPLACE the correct action to take when you come across another row with the same key? Is keeping only the 'last' row for each key what you really want to do?

Tell us more about your schema. You may not need the unique constraint (or dropping it might make your data more useful as you'll be keeping a history of all the records you've uploaded.)