So, what you want is to have the same exact results set as if the query was "scientific", but the documents that also match Field1:[20 TO 30] to have more score, right?
On Wed, Jul 20, 2011 at 10:53 AM, Sowmya V.B. <vbsow...@gmail.com> wrote: > Hi Tomas > > Here is what I was trying to give. > > > http://localhost:8085/apache-solr-3.3.0/select?indent=on&version=2.2&defType=dismax&q=scientific&bq=Field1:[20%20TO%2030] > ^10&start=0&rows=30&qf=text&fl=Field1,docid&debugQuery=on > This query seems OK for that purpose. > > Over here, I was trying to change the range of Field1, keeping everything > else intact. Here are my observations: > > 1) The number of results found remain intact. Only that the order of the > results varies. > Isn't this what was expected? > 2) The boost factor (10) does not seem to throw any influence at all. > It's on the parsed query, Why do you think it doesn't have an influence? Can you send the debug query output for a document that match the bt? I tried it with the Solr example and this is what I see: http://localhost:8983/solr/select?defType=dismax&q=display&bq=weight:[0%20TO%2010] ^10&start=0&rows=30&debugQuery=on&qf=features%20name This is the debug output for a document that match the query and the boost query: <str name="MA147LL/A"> 1.137027 = (MATCH) sum of: 0.1994111 = (MATCH) max of: 0.1994111 = (MATCH) weight(features:display in 0), product of: 0.34767273 = queryWeight(features:display), product of: 3.7080503 = idf(docFreq=1, maxDocs=30) 0.0937616 = queryNorm 0.57355976 = (MATCH) fieldWeight(features:display in 0), product of: 1.4142135 = tf(termFreq(features:display)=2) 3.7080503 = idf(docFreq=1, maxDocs=30) 0.109375 = fieldNorm(field=features, doc=0) 0.937616 = (MATCH) ConstantScore(weight:[0.0 TO 10.0]^10.0)^10.0, product of: 10.0 = boost 0.0937616 = queryNorm </str> and this is the debug output for a document that only match the main query: <str name="VA902B"> 0.4834455 = (MATCH) sum of: 0.4834455 = (MATCH) max of: 0.4834455 = (MATCH) weight(name:display in 12), product of: 0.34767273 = queryWeight(name:display), product of: 3.7080503 = idf(docFreq=1, maxDocs=30) 0.0937616 = queryNorm 1.3905189 = (MATCH) fieldWeight(name:display in 12), product of: 1.0 = tf(termFreq(name:display)=1) 3.7080503 = idf(docFreq=1, maxDocs=30) 0.375 = fieldNorm(field=name, doc=12) </str> Do you have something similar?? > Here is what the debugQuery says: > <str name="parsedquery">+DisjunctionMaxQuery((text:scientif)) () > Field1:[20.0 TO 30.0]^10.0</str> > <str name="parsedquery_toString">+(text:scientif) () Field1:[20.0 TO > 30.0]^10.0</str> > > From these, it seems like its just filtering the results based on the > Field1 > values, rather than performing a Boost Query. > > S. > > 2011/7/20 Tomás Fernández Löbbe <tomasflo...@gmail.com> > > > Yes, it should, but make sure you specify at least the "qf" parameter > for > > dismax. You can activate debugQuery and you'll see which documents get > > boosted and which aren't. > > > > On Wed, Jul 20, 2011 at 9:21 AM, Sowmya V.B. <vbsow...@gmail.com> wrote: > > > > > Hi Tomasso > > > > > > Thanks for a quick response. > > > > > > So, if I say: > > > http://localhost:8085/apache-solr-3.3.0/select?indent=on&version=2.2* > > > > &defType=dismax*&q=scientific&bq=Field1:[20%20TO%2025]^10&start=0&rows=30 > > > -will it be right? > > > > > > The above query: boosts the documents which suit the given query > > > ("scientific"), which has Field1 values between 20-25, by a factor of > 10 > > : > > > Is that right?? > > > > > > S > > > > > > 2011/7/20 Tomás Fernández Löbbe <tomasflo...@gmail.com> > > > > > > > Hi Sowmya, "bq" is a great way of boosting, but you have to be using > > the > > > > Dismax Query Parser or the Extended Dismax (edismax) query parser, it > > > > doesn't work with the Lucene Query Parser. If you can use any of > those, > > > > then > > > > that's the solution. If you need to use the Lucene Query Parser, for > a > > > user > > > > query like: > > > > > > > > scientific temper > > > > > > > > you could create a query like: > > > > > > > > (scientific temper) OR (scientific temper AND (field1:[10 TO > 2030]))^X > > > > > > > > being "X" the boost you want for those documents. > > > > > > > > with your query: > > > > scientific temper field1:[10 TO 2030] > > > > > > > > you are either adding the condition of the range value for the field > > (if > > > > your default operator is AND) or adding another way of matching the > > query > > > > (if your default operator ir OR, you can have documents in your > result > > > set > > > > that only matched the range query, and this is not what the user > > wanted). > > > > > > > > Hope this helps, > > > > > > > > Tomás > > > > > > > > On Wed, Jul 20, 2011 at 5:15 AM, Sowmya V.B. <vbsow...@gmail.com> > > wrote: > > > > > > > > > Can anyone throw some light on this issue? > > > > > > > > > > My problem is to: give a query time boost to certain documents, > which > > > > have > > > > > a > > > > > field, say field1, in the range that the user chooses during query > > > time. > > > > I > > > > > think the below link indicates a range query: > > > > > > > > > > > > > > > > > > > > > > > > > http://localhost:8085/solr/select?indent=on&version=2.2&q=scientific+temper+field1:[10%20TO%2030]&start=0&rows=10 > > > > > > > > > > But, apart from that, how can I indicate a boost for the condition > > > > > field1:[10%20TO%2030]? > > > > > > > > > > I tried using a &bq=field1:[20 TO 25] and also &bq=field1:[20 TO > > 25]^10 > > > > > -But I am not able to figure out what these two mean, from the > > results. > > > > > Because, i get top1 result as a document where field1 is 40..in > this > > > > > case..after using &bq clause. I increased the boost to 10,20,50 > > > 100..but > > > > > the > > > > > results dont change at all. > > > > > > > > > > S. > > > > > > > > > > On Tue, Jul 19, 2011 at 4:28 PM, Sowmya V.B. <vbsow...@gmail.com> > > > wrote: > > > > > > > > > > > Hi > > > > > > > > > > > > Is query time boosting possible in Solr? > > > > > > > > > > > > Here is what I want to do: I want to boost the ranking of certain > > > > > > documents, which have their relevant field values, in a > particular > > > > range > > > > > > (selected by user at query time)... > > > > > > > > > > > > when I do something like: > > > > > > > > > > > > > > > > > > > > > > > > > > > http://localhost:8085/solr/select?indent=on&version=2.2&q=scientific+temper&fq=field1:[10%20TO%2030]&start=0&rows=10 > > > > > > -I guess, it is just a filter over the normal results and not > > exactly > > > a > > > > > > query. > > > > > > > > > > > > I tried giving this: > > > > > > > > > > > > > > > > > > > > > > > > > > > http://localhost:8085/solr/select?indent=on&version=2.2&q=scientific+temper+field1:[10%20TO%2030]&start=0&rows=10 > > > > > > -This still worked and gave me different results. But, I did not > > > quite > > > > > > understand what this second query meant. Does it mean: "Rank > those > > > > > documents > > > > > > with field1 value in 10-30 better than those without" ? > > > > > > > > > > > > S > > > > > > -- > > > > > > Sowmya V.B. > > > > > > ---------------------------------------------------- > > > > > > Losing optimism is blasphemy! > > > > > > http://vbsowmya.wordpress.com > > > > > > ---------------------------------------------------- > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > Sowmya V.B. > > > > > ---------------------------------------------------- > > > > > Losing optimism is blasphemy! > > > > > http://vbsowmya.wordpress.com > > > > > ---------------------------------------------------- > > > > > > > > > > > > > > > > > > > > > -- > > > Sowmya V.B. > > > ---------------------------------------------------- > > > Losing optimism is blasphemy! > > > http://vbsowmya.wordpress.com > > > ---------------------------------------------------- > > > > > > > > > -- > Sowmya V.B. > ---------------------------------------------------- > Losing optimism is blasphemy! > http://vbsowmya.wordpress.com > ---------------------------------------------------- >
