Hi Shahzad, Personally I am of the same opinion as others who have replied, that you are better off going back to your clients at this stage itself, with all the new found info/data points.
Further, to the questions that you put to me directly: 1) For option 1, as indicated earlier, you have to compute the myfieldwordcount outside of Solr & push it in as any other field to Solr. As far as I know, there is no filter that will do this for you out of the box. 2) For option 2, you had to take a look at: http://wiki.apache.org/solr/SolrRelevancyFAQ#index-time_boosts Related links: Function Query: http://wiki.apache.org/solr/FunctionQuery#norm Norms: http://lucene.apache.org/core/3_5_0/api/all/org/apache/lucene/search/Similarity.html#computeNorm(java.lang.String, org.apache.lucene.index.FieldInvertState) Changes to schema: http://wiki.apache.org/solr/SchemaXml#Common_field_options (omitNorms option) For a field with default boost (= 1), norm = lengthNorm (approximately 1/sqrrt(numTerms)). Norm's been multiplied twice in the query to divide the score (approx.) by numTerms. Hope that helps. Regards, Aloke On Fri, Oct 11, 2013 at 5:36 PM, shahzad73 <shahzad...@yahoo.com> wrote: > Aloke Ghoshal i'm trying to work out your equation. i am using standard > scheme provided by nutch for solr and not aware of how to calculate > myfieldwordcount in first query. no idea where this count will come > from. is there any filter that will store number of tokens generated for > a > specific field and store it as another field. that way we can use it . > not sure what norm does in second equation try to find information for > this from online and did not find any yet. please explain > > > Shahzad > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Find-documents-that-are-composed-of-words-tp4094264p4094955.html > Sent from the Solr - User mailing list archive at Nabble.com. >
