That's a good point. What is the query sorting on? Shayan, can you give an example of a query with sorting/etc shown.
Regards, Alex. ---- Solr Analyzers, Tokenizers, Filters, URPs and even a newsletter: http://www.solr-start.com/ On 3 September 2015 at 16:24, Chris Hostetter <hossman_luc...@fucit.org> wrote: > > : Write a PostFilter which takes in a document id. It lets through all > : documents until it sees that document id. Once it sees it, it stops > : letting them through. > : > : Thus, the total count of documents would be the position of your queried > : car. > > Sorry guys, that won't work. > > PostFilter's can be used to collect & filter the documents returned as the > result of a query, after the main query logic (so you can delay expensive > filter checks) but they still happen before any sorting -- they have to in > order to in order for the sorting logic to know *whic* documents > should be added to the priority queue. > > - - - > > I can only think of two appraoches to this general problem: > > 1) 2 queries with frange filter on score. > > this solution is only applicable in situations where: > a) you are only sorting on scores > b) the position information can be aproximate as far as other docs with > identical scores (ie: you can say "X documents have a higher score" > instead of "exactly X documents come before this one") > > The key is to first do a query on whever where you filter (fq) on the doc > id(s) you are interested in so you can get them back along with the > scores, then you do another query where you do something like... > > ?rows=0&q=whatever&fq={!frange incl=false l=THE_SCORE v=$q} > > ...so that you post filter and ignore any doc that doesn't have a higher > score and look at hte total numFound. > > if there are multiple docs you need to get info about at one time, instead > of filtering you can use facet.query the same way > > rows=0 > q=whatever > facet=true > facet.query={!frange key=doc1 incl=false l=DOC1_SCORE v=$q} > facet.query={!frange key=doc2 incl=false l=DOC2_SCORE v=$q} > facet.query={!frange key=doc3 incl=false l=DOC3_SCORE v=$q} > ...etc... > > https://cwiki.apache.org/confluence/display/solr/Other+Parsers#OtherParsers-FunctionRangeQueryParser > > 2) cursor deep paging > > this solution will work regardless of the number of docs you are > interested in, and regardless of how complex your sort options are -- just > use the cursorMark param to iterate over all the results in your client > until you've found all the unqiueKeys you are looking for, counting the > docs found as you go. > > The various docs on deep paging and using cursors go into some background > hwich may help you understand why what you are asking for in general is a > hard problem, and why suggestion #1 only works with a simple sort on > score, and for anything more complex you really have to go the cursor > route... > > https://cwiki.apache.org/confluence/display/solr/Pagination+of+Results > https://lucidworks.com/blog/coming-soon-to-solr-efficient-cursor-based-iteration-of-large-result-sets/ > > -Hoss > http://www.lucidworks.com/
