Elasticsearch: search against _all field with boosting

Question

I am new to Elasticsearch.

I have the following mapping:

{
    "mappings": {

        "book": {

            "properties": {         
                "title": {
                    "properties": {
                        "en": {
                            "type": "string",
                            "analyzer": "standard"
                        },
                        "ar": {
                            "type": "string",
                            "analyzer": "standard"
                        }
                    }
                },

                "keyword": {
                    "properties": {
                        "en": {
                            "type": "string",
                            "analyzer": "standard"
                        },
                        "ar": {
                            "type": "string",
                            "analyzer": "standard"
                        }
                    }
                }
            }
        }
    }
}

A sample document may have two languages for the same field of the same book. Here are two example documents:

{
    "title" : {
        "en": "hello",
        "ar": "مرحبا"
    },
    "keyword" : {
        "en": "world",
        "ar": "عالم"
    }   
}

{
    "title" : {
        "en": "Elasticsearch"
    },
    "keyword" : {
        "en": "full-text index"
    }   
}

Now I want to do a search against the _all field. Here is my query:

"query": { 
    "match" : {
        "_all" : {
            "query" : "hello",
            "operator" : "OR"
        }
    }
}

Is this correct mapping? One of the reasons that I want to use _all field instead of listing the specific fields in query is that I will include more languages.

The thing I am not sure about is how to add boost to the title.en, title.ar fields in the above query? Any better way of doing so in case of more languages?

Thanks and regards.

Answer 1

You can do that by using function_score query

{
   "query": {
      "function_score": {
         "functions": [
            {
               "boost_factor": "500",
               "filter": {
                  "term": {
                     "title.en": "hello"
                  }
               }
            },
            {
               "boost_factor": "200",
               "filter": {
                  "term": {
                     "title.ar": "hello"
                  }
               }
            }
         ],
         "query": {
            "match": {
               "_all": {
                  "query": "hello",
                  "operator": "OR"
               }
            }
         },
         "score_mode": "sum"
      }
   }
}

adding title.*

{
   "query": {
      "function_score": {
         "functions": [
            {
               "boost_factor": "500",
               "filter": {
                   "query": {
                      "query_string": {
                           "default_field": "title.*",
                            "query": "hello"
                       }
                   }
               }
            }
         ],
         "query": {
            "match": {
               "_all": {
                  "query": "hello",
                  "operator": "OR"
               }
            }
         },
         "score_mode": "sum"
      }
   }
}

Answer 2

You should add this in the mapping:

_all" : {"enabled" : true}

For an example mapping check this.

You can use _all field in your case as it is suitable for your requirement. ElasticSearch Documentation states that:

The idea of the _all field is that it includes the text of one or more other fields within the document indexed. It can come very handy especially for search requests, where we want to execute a search query against the content of a document, without knowing which fields to search on. This comes at the expense of CPU cycles and index size.

You are using a match query. It will work only if there is an exact match in _all field. By default, _all field is analyzed using standard analyzer. so a search for hello world using a match might won't return a hit. I'll advise in that case using a query_string or multi_match is much better. You can specify a custom analyser for _all field like:

"_all" : {"type" : "string", "analyzer" : "your_custom_analyzer"}

Thanks