Hi Ashish, are we talking about Analysis at query or Index time or both ? As Erick say I found really hard to believe for this combination in a classic search.
Are you trying to provide something "special" ? Ngram token filter will produce a setof ngram out of your token: token to ok ke en in case of bigrams. I find this as useless input for a stemmer. Inverting the two token filter will probably make more sense. But , can we know which kind of search you want to provide on top of this analysis ? Analysis must always go in pair with the expected Search you want! Cheers 2015-07-29 10:49 GMT+01:00 Ashish Mukherjee <ashish.mukher...@gmail.com>: > Hello, > > I am using Stemmer on a Ngram field. I am getting better results with > Stemmer factory after Ngram, but I was wondering what is the recommended > practice when using Stemmer on Ngram field? > > Regards, > Ashish > -- -------------------------- Benedetti Alessandro Visiting card - http://about.me/alessandro_benedetti Blog - http://alexbenedetti.blogspot.co.uk "Tyger, tyger burning bright In the forests of the night, What immortal hand or eye Could frame thy fearful symmetry?" William Blake - Songs of Experience -1794 England
