Damn...I was obfuscated seeing the 14 there...I had naively thought that term freq would not be stored in the doc, 1 would be stored, but I guess it still stores the real value and then applies custom similarity at query time.
That means changing to a custom similarity does not need reindexing right? thanks for the help! xavier On Thu, Mar 21, 2013 at 5:26 PM, Chris Hostetter <hossman_luc...@fucit.org>wrote: > : > public class NoTfSimilarity extends DefaultSimilarity { > : > public float tf(float freq) { > : > return freq > 0 ? 1.0f : 0.0f; > : > } > : > } > ... > > : > But I still see tf=14 in my query?? > ... > : > 1.0 = tf(freq=14.0), with freq of: > : > 14.0 = termFreq=14.0 > > pretty sure you are looking at the explanation of the *input* to your tf() > function, not that the *output* is 1.0, just like in your function. > > Did you compare this to what you see using the DefaultSimilarity? > > > > -Hoss >