It is not at all clear what you are talking about.
What "alternate records" do you mean? If you're referring to alternate indexes (not records), are you wanting to unload the alternate index records themselves (which consist of a 4-byte overhead area, the alternate index key, and 4-byte pointers to each primary record with the matching alternate index), or unload the base cluster in alternate index key order?
What are you having problems with? Being specific in forum posts gets you much better answers than generic statements, since the more specific you make things the clearer it is for those of us who have no idea about your application, your site, or your requirements to help you.