Well you are correct Erik that this is a database-ish thing try to achieve in solr and unfortunately the sin :) had been committed by somebody else :) and now we are running into relevancy issues.
Let me try to state the problem more casually. 1. There are user records of type A, B, C etc. (userId field in index is common to all records) 2. A user can have any number of A, B, C etc (e.g. think of A being a language then user can know many languages like french, english, german etc) 3. Records are currently stored as a document in index. 4. A given query can match multiple records for the user 5. If for a user more records are matched (e.g. if he knows both french and german) then he is more relevant and should come top in UI. This is the reason I wanted to add lucene scores assuming the greater score means more relevance. Hope you got what I was saying. Another idea for this situation is doing faceting on userId field and then add the score but currently I think lucene only support facet count, basically solr will give you only count of docs it matched. Can I get sum of the score of documents that matched? Thanks Bharat Jain On Tue, Jul 27, 2010 at 5:58 AM, Erick Erickson <erickerick...@gmail.com>wrote: > I'm having trouble getting my head around what you're trying to accomplish, > so if this is off base you know why <G>. > > But what it smells like is that you're trying to do database-ish things in > a SOLR index, which is almost always the wrong approach. Is there a > way to index redundant data with each document so all you have to do > to get the "relevant" users is a simple query? > > Adding scores is also suspect.. I don't see how that does predictable > things. > > But I'm also failing completely to understand what a "relevant" user is. > > not much help, if this is way off base perhaps you could provide some > additional use-cases? > > Best > Erick > > On Mon, Jul 26, 2010 at 2:37 AM, Bharat Jain <bharat.j...@gmail.com> > wrote: > > > Hello All, > > > > I have a index which store multiple objects belonging to a user > > > > for e.g. > > <schema> > > <field name="objType" type="..." /> ---------> Identifies user > > object type e.g. userBasic or userAdv > > > > <!-- obj 1 --> > > <field name="first_name" type="..." /> ----> MAPS to > userBasicInfoObject > > <field name="last_name" type="..." /> > > > > <!-- obj 2 --> > > <field name="user_data_1" type="..." /> -----> MAPS to userAdvInfoObject > > <field name="user_data_2" type="..." /> > > > > </schema> > > > > > > Now when I am doing some query I get multiple records mapping to java > > objects (identified by objType) that belong to the same user. > > > > > > Now I want to show the relevant users at the top of the list. I am > thinking > > of adding the Lucene scores of different result documents to get the best > > scores. Is this correct approach to get the relevance of the user? > > > > Thanks > > Bharat Jain > > >
