I have a document library in which I need to crawl the columns within that library but not crawl the text within the documents themselves. The documents are OCR'd pdf's from a scanner. Is there a way to set this up?
By default SP won't index PDF unless you installed an iFilter to do so. So, if you don't have an iFilter, no need to worry, it won't get indexed.
On the other hand if you do have an iFilter for PDF indexing, you are in for some work to exclude only parts of a crawl, there are no "settings" you can configure for this type of scenario.
The only work around I can readily think of (and this is theory) would be to convert from a document library to a list with a document attachment, then in Central Administration exclude the document URL from search results. Anything containing that URL in search will be removed.
Since columns in a doc library will be linked with that document, the doc link would be a part of search results and thus removed from all indexes including your column entries.
As a list attachment it's just another column item and you would only be excluding that column link, not the others, so you should still be able to search on the other column data.
Outside of this, you will probably need to dig into the Search Object Model for a solution. For a large amount of items this is a lot of work, and if it needs to continually be updated, you probably need to look at automating it via the object model anyway.
If it didn't, help I hope it gave you some ideas.