Thanks, the solution is now detailed very clearly.
Just one more point, more theoretical then technical: i'm using Solr in a e-commerce site, and i wanted to use protected words also to reduce recall for certain queries. It could be that a parallel approach using dismax boosting for fields such as "product name" and "category" will, beside increasing precision, also reducing false hit recall? hossman wrote: > > > : I have a question about using the "protected=" attribute with > : SnowballPorterFilterFactory filter. > > SnowballPorterFilterFactory doesn't (and has never) supported a protwords > option ... that feature is unique to the EnglishPorterFilterFactory. > > this is probably just due to how they came about back in the pre-Apache > days of Solr ... the lucene SnowballFilter has to use reflection because > of the way the underlying Snowball Stemmer API works, and the > SnowballPorterFilterFactory just generates intsances of SnowballFilter. > Yonik wrote EnglishPorterFilterFactory to be more efficient by not doing > the reflection (at the expense of being hardcoded to use English) and when > the protwords feature was needed later it looks like it was only added > there. > > adding a protwords option to SnowballPorterFilterFactory would be > possible, but the best way to do it would probably be to add support in > the underlying Lucene-Java class, then add a config option for it in Solr. > > (if anyone is so inclined) > > > > > -Hoss > > > -- View this message in context: http://www.nabble.com/SnowballPorterFilterFactory-and-protected-words-tp15042758p15061335.html Sent from the Solr - User mailing list archive at Nabble.com.
