Hi Pragyanshis, I faced a similar problem few days ago and I was advised on this forum to override Solr DefaultSimilairy calculation to return always a constant value for idf. I think, in your case you'd also want to suppress the length norm which will require re-indexing as length norm is calculated during indexing.
The link of my issue is as below: http://lucene.472066.n3.nabble.com/Possible-issue-in-edismax-td4037397.html Cheers, Sandeep On 14 February 2013 19:20, Pragyanshis Pattanaik <pragyans...@outlook.com>wrote: > Hi, > My schema is like below. > <fields> <dynamicField name="Subject-Name-*" type="string" > indexed="true" stored="true"/> <dynamicField > name="Subject-Mark-*" type="int" indexed="true" stored="true"/> </fields> > My need is to search only three subject fields and boost those subjects > which has a higher Mark(Mark can be in between 1 - 10). > Again Top subjects will get a higher boost than preceding one's. > Like if a search term is present in Subject-Name-1,Then it will get a > higher boost than Subject-Name-2 and Subject-Name-3. > Similarly Subject-Mark-1 will get higher boost than Subject-Mark-2 and > Subject-Mark-3. > To achieve this i am querying over subject fields and my query looks like > below. > > q=+Economics+Geography&wt=xml&deftype=edismax&qf=Subject-Name-1+Subject-Name-2+Subject-Name-3&bq=Subject-Name-1%3AEconomics%3BGeography^50.0+Subject-Mark-1%3A20^90.0+Subject-Mark-1%3A9^80.0+Subject-Mark-1%3A8^70.0+Subject-Mark-1%3A7^60.0+Subject-Name-2%3AEconomics%3BGeography^45.0+Subject-Mark-2%3A20^90.0+Subject-Mark-2%3A9^80.0+Subject-Mark-2%3A8^70.0+Subject-Mark-2%3A7^60.0+Subject-Name-3%3AEconomics%3BGeography^40.0+Subject-Mark-3%3A20^90.0+Subject-Mark-3%3A9^80.0+Subject-Mark-3%3A8^70.0+Subject-Mark-3%3A7^60.0 > If i am having four documents like below > <doc> <str name="Subject-Name-1">Economics</str> <str > name="Subject-Name-1">Geography</str> <str > name="Subject-Name-1">History</str> <int name="Subject-Name-1">7</int> > <int name="Subject-Name-1">7</int> <int name="Subject-Name-1">6</int> > </doc> <doc> <str name="Subject-Name-1">Economics</str> <str > name="Subject-Name-1">History</str> <str > name="Subject-Name-1">Geography</str> <int name="Subject-Name-1">8</int> > <int name="Subject-Name-1">8</int> <int name="Subject-Name-1">5</int> > </doc> <doc> <str name="Subject-Name-1">Economics</str> > <str name="Subject-Name-1">History</str> <str > name="Subject-Name-1">Geography</str> <int name="Subject-Name-1">9</int> > <int name="Subject-Name-1">6</int> <int name="Subject-Name-1">7</int> > </doc> <doc> <str name="Subject-Name-1">Economics</str> > <str name="Subject-Name-1">Mathematics</str> <str > name="Subject-Name-1">History</str> <int name="Subject-Name-1">7</int> > <int name="Subject-Name-1">7</int> <int name="Subject-Name-1">6</int> > </doc> > > then i am getting a higher score for last document which has only one of > the search term !!! > But in my situation it is not applicable. My requirement is,if a document > has only one term then they should get a lower score than the documents > which are having both of the terms. > Is it happening because of idf(rarer terms give higher contribution to the > total score) ? > Or there is something wrong with my query ? > Can anybody help me to achieve the desired output. > Thanks in advance
