Hi everyone,
Sorry for the n00b post. I just started using ElasticSearch for a
dynamic website search. In English, everything just works great.
Love the json interface- was able to create an index easily, query it
easily, and integrate it into the website easily.
It went so well, that I've been asked to use ElasticSearch for a
Japanese text website search. Should be no problem, the documentation
says Lucene's "cjk" analyzer is supported.
So how do I do it? Everything's in utf8, I try to execute a query:
curl -XGET 'http://localhost:9200/MySite/search/_search?pretty=1&q=
\uff52\uff49\uff50'
This fails to return any results. Naturally- I need the cjk analyzer
to break out the Japanese characters. Cjk will do it:
curl 'localhost:9200/MySite/_analyze?pretty=1&analyzer=cjk' -d '
\uff52\uff49\uff50'
And this breaks out the tokens without a problem.
Doing what I thought was the obvious:
curl 'localhost:9200/MySite/search/_search?pretty=1&q=
\uff52\uff49\uff50' -d {
"settings":{
"analysis":{
"analyzer":"cjk"
}
}
}'
This gives an "unable to parse" error, so obviously, I've done
something really stupid with my syntax.
Hence my n00b question: what's wrong with my syntax? What dumb thing
have I done? Do I need to set some default elsewhere to use the cjk
analyzer?
Thanks in advance,
-Nathan