Custom analyzer not working in elasticsearch

Question

Running elastic version 1.6

I am trying to set custom analyzer for my index in elasticsearch. My index /has some properties which contains some accents and special characters.

Like one of my property name has value like this, "name" => "Está loca". So what I want to achieve is, whenever I am trying to search by this way, http://localhost:9200/tutorial/helloworld/_search?q=esta

I should get the result for "Está loca". I have gone through following link and configured necessary analyzer which is explain in the link. https://www.elastic.co/guide/en/elasticsearch/guide/current/asciifolding-token-filter.html

curl -XPUT 'localhost:9200/tutorial?pretty' -H 'Content-Type: application/json' -d'
{
"mappings":{
  "helloworld":{
  "properties": {
    "name": { 
      "type":           "string",
      "analyzer":       "standard",
      "fields": {
        "folded": { 
          "type":       "string",
          "analyzer":   "folding"
        }
      }
    }
  }
}
},
"settings": {
    "analysis": {
      "analyzer": {
        "folding": {
          "tokenizer": "standard",
          "filter":  [ "lowercase", "asciifolding" ]
        }
      }
    }
  }
}'

I have configured this while creating index and made some entries like this for test,

curl -X POST 'http://localhost:9200/tutorial/helloworld/1' -d '{ "name": "Está loca!" }'
curl -X POST 'http://localhost:9200/tutorial/helloworld/2' -d '{ "name": "Está locá!" }'

but while searching like this, http://localhost:9200/tutorial/helloworld/_search?q=esta nothing is happening. I just want whenever a user searches in any languages for example in English it should get the same result. Please anybody can help, how can I achieve this struggling on it for last 1 week.

Answer 1

you would not be able to search for esta keyword in _all field. As elasticsearch by default only apply standard analyzer while constructing _all field.

so your following query

GET folding_index1/helloworld/_search?q=esta

Produces following match query in elastic dsl.

GET folding_index1/helloworld/_search
{
  "query": {
    "match": {
      "_all": "esta"
    }
  }
}

Which search against _all field and hence couldn't find folded token for name.

You can do following, but even with include_in_all mentioned for multifield, it still applies standard analyzer for _all field.

PUT folding_index1
{
    "mappings": {
        "helloworld": {
            "properties": {
                "name": {
                    "type": "string",
                    "analyzer": "standard",
                    "fields": {
                        "folded": {
                            "type": "string",
                            "analyzer": "folding",
                            "include_in_all": true
                        }
                    }
                }
            }
        }
    },
    "settings": {
        "analysis": {
            "analyzer": {
                "folding": {
                    "tokenizer": "standard",
                    "filter": ["lowercase", "asciifolding"]
                }
            }
        }
    }
}

Query like following can work for you. More on _all field analyzer

POST folding_index1/_search?q=name.folded:esta