ElasticSearch - prefix with space and filtering

Question

My ElasticSearch server contains documents of the following form:

{
    "_index": "xindex",
    "_type": "xtype",
    "_id": "1100",
    "_score": 3.00010,
    "_source": {
        "_id": "2333345",
        "field1": "11111111111111",
        "field2": "y",
        "name": "hello world",
    }
}

I need to get all the documents with name prefix "hello wo" and field2 "y". Tried a lot of queries and none have worked. There are all kind of solutions for the prefix with space issue, but when adding the filtering/another query for field2, results get corrupted.

Thanks.

Answer 1

• use simple_query_string, This approach solved my issue:

{
    "query": {
        "bool": {
            "should": [
                {
                    "simple_query_string": {
                        "fields": [
                            "name"
                        ],
                        "default_operator": "and",
                        "query": "(hello world*)"
                    }
                }
            ]
        }
    }
}

Answer 2

You can achieve this in 3 steps :

1. Change your mapping of field name to not_analyzed
2. Use a match_phrase_prefix query (documentation here)
3. Filter this query results by wrapping it in a filtered query and use a term filter on the field2 with value "y"

You can see it working with the following dataset :

PUT test/prefix/_mapping
{
  "properties": {
    "name":{
      "type": "string",
      "index": "not_analyzed"
    }
  }
}

//should match
PUT test/prefix/2333345
{
  "field1": "11111111111111",
  "field2": "y",
  "name": "hello world"
}

//should match    
PUT test/prefix/1112223
{
  "field1": "22222222222222",
  "field2": "y",
  "name": "hello wombat"
}

//should not match (field2 value is different)
PUT test/prefix/4445556
{
  "field1": "33333333333333",
  "field2": "z",
  "name": "hello world"
}

//should not match (second word not starting with wo)
PUT test/prefix/4445556
{
  "field1": "33333333333333",
  "field2": "y",
  "name": "hello zombie"
}

Then, the query is :

GET test/prefix/_search
{
  "query": {
    "filtered": {
      "query": {
        "match_phrase_prefix" : {
        "name" : "hello wo"
        }
      },
      "filter": {
        "term": {
          "field2": "y"
        }  
      }
    } 
  }
}

which outputs the documents 1112223 and 2333345 as expected :

{
   "took": 20,
   "timed_out": false,
   "_shards": {
      "total": 5,
      "successful": 5,
      "failed": 0
   },
   "hits": {
      "total": 2,
      "max_score": 1.592944,
      "hits": [
         {
            "_index": "test",
            "_type": "prefix",
            "_id": "2333345",
            "_score": 1.592944,
            "_source": {
               "field1": "11111111111111",
               "field2": "y",
               "name": "hello world"
            }
         },
         {
            "_index": "test",
            "_type": "prefix",
            "_id": "1112223",
            "_score": 1.592944,
            "_source": {
               "field1": "22222222222222",
               "field2": "y",
               "name": "hello wombat"
            }
         }
      ]
   }
}