Sounds like you are tokenizing your string when you don't really want to. Either you want all queries to only search against prefixes of the whole value without tokenization, or you need to produce several copyFields with different analysis applied and use dismax to let Solr know which should rank higher.
Or, you could use the Suggester component or one of the other bolt-on autocomplete components instead. Maybe you should post your current field definition and let us know specifically what you're trying to achieve? Michael Della Bitta Applications Developer o: +1 646 532 3062 appinions inc. “The Science of Influence Marketing” 18 East 41st Street New York, NY 10017 t: @appinions <https://twitter.com/Appinions> | g+: plus.google.com/appinions <https://plus.google.com/u/0/b/112002776285509593336/112002776285509593336/posts> w: appinions.com <http://www.appinions.com/> On Thu, May 29, 2014 at 4:54 AM, Gurfan <htc.ja...@gmail.com> wrote: > Hi All, > > We are using EdgeNGramFilterFactory for searching with minGramSize="3", as > per Business logic, auto fill suggestions should appear on entering 3 > characters in search filter. While searching for contact with name "Bill > Moor", the value will does not get listed when we type 'Bill M' but when > we type 'Bill Moo' or 'Bill' it suggests 'Bill Moor'. > > Clearly, The tokens are not generated when there is space in between, we > cannot set set minGramSize="1" as that will generate many tokens and slow > the performance. Do we have a solution without using Ngram to generate > tokens on entering 3 characters? > > > Please suggest. > > Thanks, > --Gurfan > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/search-using-Ngram-tp4138596.html > Sent from the Solr - User mailing list archive at Nabble.com. >