COPY optimization issue

My postgres database does a nightly sync with a Legacy database by
clobbering a postgres table with data from a CSV.

I currently just use the COPY command, but with over 800,000 records, this
takes quite some time.

Is there a faster way?

eg I notice that any validation failure of ANY records causes the entire
copy to roll back. Is this begin/commit action wrapped around the copy
costing me CPU cycles? And if so, can I turn it off, or is there a better
way then using copy?

Note: I do nightly vacuum's so deleted tuples is not the issue, I don't
think.

Thanks

Terry Fielder
Network Engineer
Great Gulf Homes / Ashton Woods Homes
terry@greatgulfhomes.com