We have a client using Sharepoint 2007 Enterprise search to crawl a load of documents on a file share. They want to associate additional properties with each crawled document. (These properties come from a separate store - e.g. given the path to a file, I can retrieve the properties for that file.) Then we can map these crawled properties to managed properties.
My first thought was to use BDC to access the properties for each file and then index this, but this loses the link between the file properties and the full-text search of the files themselves.
I've come across this relevant question on SO to which the response is "use a custom protocol handler". This sounds good, but I still don't know if this can hook into the regular full-text search, or if we'd have to reimplement the full-text search in the custom handler.
Has anyone implemented a custom protocol handler and can tell me whether it can be used alongside the regular full-text search?
Alternatively, can anyone suggest a different method for associating custom properties with files in a crawled file share?