: As I mentioned previously, I prefer to do this with as little java : code as possible. That's the motivation for me to take a look at solr.
I understand, but as i already said "there is no pure configuration way to obtain the same logic you could get from a custom HitCollector" you can get the same behavior you currently have, with the same existing efficiencies plus take advantage of the solr filter cache by writting writing a custom RequestHandler (or SearchCOmponent) that would be about 5 lines long... get the DocSet from the searcher for your parsed query (you can reuse the existing Solr QueryParser framework and utilities) then iterate over the DocSet and add each FieldCache value to the response. FWIW: I would encourage you to try using Solr as is, w/o any custom code or messing with the field cache and just set "fl=yourField" and see if the performance is satisfactory to you. it will still do scoring but you might be suprised how fast stored fields can be returned (under the covers solr uses a FieldSelector contain just the "fl" fields) -Hoss
