It turns out that index was not updated for some reason and then it worked better but still something was wrong.
I am sorry for the inconvenience. Moreover I must used other search_analyzer as opposed to analyzer. Analyzer is ngram and search_analyzer is keyword.
Having ngram in both search_analyzer and analyzer I searched "rk tw" and it returned too many results, like Tom Twain or something like his. I guess it didn't keep the sequence in search string and instead search words with "rk tw" substring it search words having "rk" or "tw" inside.
Today I know that I can setup mapping which would index the same field in document different manner and different names and it is possible to boost fields so maybe I will play with it a bit more.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.