Re: Phrase query not matching exact tokens in some cases

2015-07-14 Thread Erick Erickson
Also, the first place to look for answers for questions like "what is the stemmer doing" is to look at the admin/analysis page. Each step in the analysis chain will be shown. Hover over the light gray initials and you'll see the class used (e.g. WST == WhitespaceTokenizer). Best, Erick On Tue, Ju

Re: Phrase query not matching exact tokens in some cases

2015-07-14 Thread Alessandro Benedetti
Which kind of Highlighter are you using ? Anyway it is responsibility of your analysis chain. it is an heavy analysis chain and I can see : "solr. HunspellStemFilterFactory" If you are using the term vector for your field, to be used by your highlighter, in the term vector , for each document, you

Phrase query not matching exact tokens in some cases

2015-07-14 Thread Mike Thomsen
For the query "police office" our users are getting back highlighted results for "police office*r*" (and "police office*rs*") I get why a search for police officers would include just "office" since the stemmer would cause that behavior. However I don't understand why "office" is matching "officer"