I always feel guilty for posting in these kinds of forums because I know my question has been answered before, I just don't know what to search to find it.

I have URLs stored in a table (essentially just an ID and the text of the URL link) and they need to be globally unique. For performance, I have a data structure that stores some instances in memory to minimize trips to the database. When I parse a web page and it has several links, I want to efficiently find all the URLs that already exist in the database, and create those that don't. Specifically:

• Check my in-memory cache of URL instances, comparing the text of the URL. Return the matching detached instances.
• Check the database for existing URL instances, detach those instances, cache them in the in-memory cache for future use, and return them. The slow way of doing this is, for each URL,
  
  Code:
  URL u = (URL) s.createCriteria(URL.class).add(Restrictions.eq("url_text", "www.example.com")).uniqueResult();
  if (u != null) {
      ...
  }
• For the remaining URLs, create new URL instances and save them in the database. Cache them in the in-memory cache for future use.

How can do this efficiently/in bulk? I'm basically coming to Hibernate with fully initialized URL POJO's except for the persistent ID, and I just need Hibernate to "fill in" the ID with existing IDs or allocate new ones. What I would like to do is similar to an "upsert", but different because I don't want to update existing instances, and that query doesn't return back any of the rows. thanks in advance for any help!

Does a anyone have any ideas or can anyone point me to where in the documentation it talks about this kind of topic? thanks