Defining analyzer per query

Elastic Stack Elasticsearch
1

Hi,
I have read in the documentation that it should be possible to specify analyzer per query. I tried a few examples with custom ngram analyzer set for match query, but it was not working as I expected.

put people
{
	"mappings": {
		"people": {
			"properties": {
				"name": {
					"type": "text"
				}
			}
		}
	},
	"settings": {
		"analysis": {
			"analyzer": {
				"ngramAn": {
					"type": "custom",
					"filter": [
						"lowercase",
						"ngramFilt"
					],
					"tokenizer": "standard"
				}
			},
			"filter": {
				"ngramFilt": {
					"type": "ngram",
					"max_gram": 4,
					"min_gram": 3
				}
			}
		}
	}
}
POST people/people/
{
    "name" : "kimchy"
}

GET people/_search
{
    "query": {
        "match" : {
            "name" : {
                "query" : "kim",
                "analyzer": "ngramAn"
            }
        }
    }
}

Have I done something wrong or is the problem in the bad understanding of analyzers (it is not clear from documentation if the query analyzer does not apply only search term - so in this case only "kim" is analyzed by ngramAn and the results of this analysis is matched against "name" field with standard analyzer).

2

Here is what you should do instead:

DELETE people
PUT people
{
	"mappings": {
		"_doc": {
			"properties": {
				"name": {
					"type": "text",
					"analyzer": "ngramAn",
					"search_analyzer": "simple"
				}
			}
		}
	},
	"settings": {
		"analysis": {
			"analyzer": {
				"ngramAn": {
					"type": "custom",
					"filter": [
						"lowercase",
						"ngramFilt"
					],
					"tokenizer": "standard"
				}
			},
			"filter": {
				"ngramFilt": {
					"type": "ngram",
					"max_gram": 4,
					"min_gram": 3
				}
			}
		}
	}
}
POST people/_doc
{
    "name" : "kimchy"
}

GET people/_search
{
    "query": {
        "match" : {
            "name" : {
                "query" : "kim"
            }
        }
    }
}
3

Yes, I know how to set custom analyzer in mapping, but I was curious about "The analyzer defined in a full-text query." mentioned in https://www.elastic.co/guide/en/elasticsearch/reference/current/analyzer.html

4

In your example, the analyzer ngramAn is applied to kim.
This is producing the following terms: [ "kim" ].

At index time you indexed kimchy which was analyzed as [ "kimchy" ].

So kim does not match kimchy. That's why you can't find the document.

Reason is that analyzer set on a query is applied to the searched terms not onto the existing index.

HTH

5

Thats what was not clear to me from documentation, thank you for your help.

6

Would you like to contribute a change in the documentation to make it more explicit?

7

I am afraid that my english level is not good enough for this :slight_smile:

8

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.