0

I asked this previously as a Regex question yet the issue appears to be Sphinx as even the correct Regex set-up failed in Sphinx. So: Is there a way to convert curly apostrophes in the index to straight ones? I have tried:

  • Exceptions.text: ’ => '
  • regexp using the actual character: regexp_filter=(\w+)\’s=>\1
  • regexp using the unicode: regexp_filter=(\w+)\x{2019}s=>\1

And nothing has worked. I get that the ’ may fail because these are text files but not sure why unicode would. In any event the ’ is breaking things for me and I'd hate to do mysql replace on the entire database just for this.

user3649739
  • 1,829
  • 2
  • 18
  • 28
  • Have you considered http://sphinxsearch.com/docs/current.html#conf-ignore-chars – barryhunter Mar 20 '17 at 19:45
  • @barryhunter I tried `ignore_chars = U+2019` and no luck. I'm still forced to manually remove. So far sphinx does not recognize that either as an explicit character `’` or as unicode `U+2019`. In an ideal world I'd be able to remap U+2019 to the standard `'` but at this point would just love to remove it in the index as my last option is to do a massive mysql replace for each `’` – user3649739 Mar 22 '17 at 13:32
  • If the regex is not working right, maybe ther data reaching indexer (when rexexp_filter is actully applied, si not actully proper UTF8 encoded? are you using SET NAMES in a sql_query_pre for example? – barryhunter Mar 23 '17 at 17:30

0 Answers0