I'm wondering if there's been any progress on an issue described a year or so ago in "More details on my MoreLikeThis mlt.qf boosting problem" <http://markmail.org/thread/nmabm5ly3wk2nqyy>, where it was pointed out that the MoreLikeThis handler only queries one field for each of the "interesting terms" that it finds in the input text.
I was hoping that using /mlt?mlt.fl=title+text&mlt.qf=title^2+text^0.5&mlt.interestingTerms=details&stream.body=tony+blair would produce title:tony^2 title:blair^2 text:tony^0.5 text:blair^0.5 but it actually produces just text:tony^0.5 text:blair^0.5 i.e. despite including the "title" field in both mlt.qf and mlt.fl, it only searches the "text" field. If I set mlt.fl=title, it produces title:tony^2 title:blair^2 so it is having an effect, just not the one I'm hoping for... As it stands, in Solr 1.4, the MoreLikeThis result set from the query above doesn't produce the document with title "Tony Blair" as the first result, which would seem appropriate given the input text "tony blair" and a boost on the title field. alf
