Hi Vitaliy, I don’t know of any combination of built-in Lucene/Solr analysis components that would do what you want, but there used to be filter called ShingleMatrixFilter that (if I understand both that filter and what you want correctly), would do what you want, following an EdgeNGramFilter: <https://lucene.apache.org/core/3_6_2/api/all/org/apache/lucene/analysis/shingle/ShingleMatrixFilter.html>
It was deprecated in v3.1 and removed in v4.0 (see <https://issues.apache.org/jira/browse/LUCENE-2920>) because it wasn’t being maintained by the original creator and nobody else understood it :). Uwe Schindler put up a patch that rewrote it and fixed some problems on <https://issues.apache.org/jira/browse/LUCENE-1391>, but that was never finished/committed. What you want could create a huge number of terms, depending on the # of documents, terms in the field, and term length. What do you want to use these terms for? Steve > On Oct 17, 2015, at 10:33 AM, vitaly bulgakov <bulgako...@yahoo.com> wrote: > > /why don't you put EdgeNGramFilter just after ShingleFilter?/ > > Because it will do Edge Ngrams over a shingle as a string: > for "Home Improvement" shingle it will do: .... Hom, Home, Home , Home I, > Home Im, Home Imp ...... > > But I need: > ... Hom Imp, Hom Impr ...... > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Tokenize-ShingleFilterFactory-results-and-apply-filters-to-tokens-tp4234574p4234872.html > Sent from the Solr - User mailing list archive at Nabble.com.
