Fuzzy lookup for query string on ElasticSearch

Question

I'm using ElasticSearch for letting users search through text fields, consisting of joined tag strings. The query looks like this and it works nicely:

{
    'query' : {
        'query_string' : {
            'query' : 'user query with +bool AND operators',
            'default_operator' : 'AND',
            'fields' : ['tag_string'],
            'analyzer' : 'my_analyzer'
        }
    }
}

However, I'd like to enable fuzzy matching so that British English and American spelling are covered. E.g. I'd like to get the same results for "gray" and "grey" or for "color" and "colour".

This can be done by the user by using the fuzzy operator "~" - so searching for "color~" matches both "color" and "colour". But that should be done automatically ... yet, the search query may contain bool operators and thus, may be complex.

paweloque · Accepted Answer · 2014-02-27 16:44:01Z

3

You can either use the fuzzy query:

{
    "fuzzy" : { "user" : "ki" }
}

Or use the fuzziness factor in a match query. Another way to achieve what you want in your example is to use synonyms. With synonyms you can tell elasticsearch to store synonyms to your words along with the original words, e.g. gray will be stored as gray and grey.

Here is a in-depth description of the synonyms: http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/analysis-synonym-tokenfilter.html

answered Feb 27, 2014 at 16:44

paweloque

19k27 gold badges87 silver badges138 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Simon Steinberger Over a year ago

Thanks! Could you explain a bit more about how the fuzziness factor works? Are those heavy requests or is performance okay with them? Synonyms may be tedious, since we're working with 20 languages :-P

manuel · Accepted Answer · 2015-09-10 10:33:30Z

2

Another example of fuzzy search (if you gonna use it)

POST /IndexName/TypeName/_search?size=200
{
   "query": {
      "fuzzy": {
         "FieldName": {
            "value": "gray",
            "fuzziness": 2,
            "prefix_length": 1,
            "boost": 5
         }
      }
   }
}

for multi word search use fuzzy_like_this

POST /IndexName/TypeName/_search?size=200
{
   "query": {
      "fuzzy_like_this": {
         "fields": ["FieldName1","FieldName2"],
         "like_text": "user query with +bool AND operators",
         "max_query_terms": 12,
         "fuzziness": 0.5
      }
   }
}

answered Sep 10, 2015 at 10:33

manuel

1,8581 gold badge16 silver badges16 bronze badges

Collectives™ on Stack Overflow

Fuzzy lookup for query string on ElasticSearch

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related