1

I have mysql innodb database which contains information collected from several sources, which are written in Serbian. Some of text contains latin characters, some not for same word.

I have problems like this

Ćuprija Cuprija

represents same word, and when I do fulltext search I must specify do I need to find Ćuprija or Cuprija, is there some settings how should I make selection for both results?

emir
  • 1,336
  • 1
  • 14
  • 33
  • Have you checked this http://stackoverflow.com/questions/1381186/fulltext-search-with-innodb?rq=1 and http://stackoverflow.com/questions/4048128/mysql-innodb-text-search-options?rq=1 – mysqlrockstar Apr 02 '15 at 11:37
  • innodb is now supporting full text search, in that articles they are searching for alternatives. – emir Apr 02 '15 at 11:42
  • Can you please tell us the collation for the columns you're searching? – O. Jones Apr 02 '15 at 11:50

1 Answers1

1

Change the column's collation to utf8_general_ci or utf8_unicode_ci

With utf8_bin, those two words are treated differently. With any _ci collation, case and accents are ignored.

Rick James
  • 135,179
  • 13
  • 127
  • 222
  • When I try to change collation to utf8_general_ci I get message #1283 - Column 'tekst' cannot be part of FULLTEXT index – emir Apr 03 '15 at 14:10
  • 1
    I have removed indexes, changed collation and recreated indexes, now it work. Thank you Rick James – emir Apr 03 '15 at 14:15