I'm using Lucene-search on a local copy of Wikipedia and the search results are missing *some* exact matches. For example, a search for 'Melbourne' does not turn up the city Melbourne (which is the top result on Wikipedia). However, other search terms have multiple exact match results. For example, a search for 'Australia' turns up the page for 'Australia' three times. I built the lucene indexes without any errors (over quite a few days) and the Melbourne page is present in the Mysql database. I'm using lucene-search-2.1.3. Any pointers much appreciated.
Topic on Extension talk:Lucene-search
The pages are stored according to their database ID, so the only thing I can think of is that somehow the XML dump is corrupted such to have wrong IDs in them. Can you verify that all metadata in the XML dump attached to those pages is correct?