Dear List, I have been studying Solr to build up an index of musical incipit encoded as strings into bibliographical record to retrofit this kind of search into an existing database. Basically we store the incipit data (filtered through a custom TokenFilter) as a multi valued field (one for each different incipit) inside a doc (which is my bibliographical record). Searching is very good and precise, but I have a problem: I cannot figure out how to know which one of the values in my multivalued field generated the hit! I only get a reference to a whole doc (and when I have 40 incipits it is a bit of a problem). I thought I could just manually cycle all the incipits in all the matching docs and match by hand, but this is not easy since the values are mangled by the TokenFilter. I saw some references that a solution to this problem is to use a Highlighter and then extract the matching value. In principle this works (I used FastVectorHighligter), but I have an additional problem: when searching very broad queries (like all the incipits which start with a C) I get obviously a big list of results (65k), but the highlighter would match only a very small subset of them (2 or 3), whilist the query would return all the correct (paginated) results. Attaching a debugger and tracing the highlighter code I found that the FieldQuery just rewrites the fist 1024 queries, and hence for all the results > 1024 it is very easy that my tokens will not get highlighted (and I cannot retrieve my value in the multivalue field). Can anyone help me out here? is there something very obvious I am missing? Is there an easy mechanism to just get the field that matched a query in a multiValued one? Thanks! Rodolfo
