--- On Fri, 2/18/11, Robert Haschart <rh...@virginia.edu> wrote:
> From: Robert Haschart <rh...@virginia.edu> > Subject: Re: solr.KeepWordsFilterFactory confusion > To: solr-user@lucene.apache.org > Date: Friday, February 18, 2011, 10:19 PM > Thanks for your response. After > making that change it seemed at first like it made no > difference, after restarting the jetty server, and > reindexing the test object, the display still shows: > > <arr name="format_facet"> > <str>Video</str> > <str>Streaming Video</str> > <str>Online</str> > <str>Gooberhead</str> > <str>Book of the Month</str> > </arr> > > But it turns out that I had been making an incorrect > assumption. I was looking at the retruned stored > values for the solr document, and seeing the "Gooberhead" > entry listed, and thinking that the analyzer wasn't > running. However as I have subsequently figured out, > the analyzers are not run on the data that is to be stored, > only on the data that is to being indexed. > So after making your change to that field type statement, > if I search > for format_facet:Gooberhead I > get results = 0 which is what I'd expect. But seeing > that the unexpected values are still stored with the solr > document, it seems that I will have to take a different > approach. Facets are populated from indexed values. However deleted documents (and their terms) are not really deleted until an optimize. Issuing an optimize may help in your case.
