Lucene Indexing Question

[George Kinsman <ge...@georgekinsman.com>]

Hi there,


I think Lucene.net might be a good fit for my current problem, but I can't quite work out how I'm going to make it work.


I have a field on my document that will contain multiple tags where each tag will have a numeric value with it, e.g. ['soap-5', 'sand-10', 'sunshine-14']. The value contained in the tag will represent its weight when ranked for queries, with the idea that a query of 'soap sand' would return a score of 15, and would be ranked above a document with just ['soap-5'].


I recently discovered function queries which I think would solve the problem quite nicely, except that splitting each of those strings at query-time would create a lot of memory pressure. I can't help but think there might be a better index-time approach to this. Could I perhaps write a custom analyzer that tokenises each of my tags and stores a score somewhere? I don't think this is quite faceted search, because I'm not filtering - it's just ranking.


Any tips or guidance is greatly appreciated!


Thanks,

George Kinsman