Thank you Robert for detailed reply On Fri, Apr 20, 2012 at 12:37 AM, Robert Muir <rcm...@gmail.com> wrote: > On Thu, Apr 19, 2012 at 7:26 AM, Vladimir Gubarkov <xon...@gmail.com> wrote: >> New analyzer: >> [aaa.bbb.com, 8888, a, b, c, d'e, f, g, h, i, j, k, l_m, n, o, p, q, >> r, s, t, u, v, z, y, z] >> Old analyzer: >> [aaa, bbb, com, 8888, a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, >> q, r, s, t, u, v, z, y, z] >> >> Please note the differences. > > Right, the tokenizer has changed. This is mentioned in the javadocs: > http://lucene.apache.org/core/3_6_0/api/contrib-analyzers/org/apache/lucene/analysis/ru/RussianAnalyzer.html > >> >> The most uncomfortable in new behaviour to me is that in past I used >> to search by subdomain like >> bbb.com:8888 >> and have displayed results with www.bbb.com:8888, aaa.bbb.com:8888 and >> so on. Now I have 0 results. > > Don't simply set your version parameter to 3.6 without reindexing. > This is really important!!!!!!!!!!! > Otherwise it defeats the whole purpose. >
Hmmm... I know this and I reindexed! I'll try to explain the problem (fortunately, already solved by using LUCENE_30) ones again: When indexing with new analyzer the whole lexeme "some.cool.site.com" goes to index, not 4 lexems "some", "cool", "site", "com". So it's now imposible to find this document with query: "site.com". I'm having an RSS subscription for that search, and now it's broken. --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org