I'm new to Elastic Search, I have document like below :

Mapping of same JSON index is like below :

Mapping

{
      "mappings": {
        "properties": {
          "age": {
            "type": "long"
          },
          "hobbiles": {
            "type": "keyword"
          }
        
      }
    }
}

Some sample documents are like below :

        [{
        "_id": "[email protected]",
        "age": 12,
        "hobbiles": [{
                "name": "Singing",
                "level": "begineer"
            },
            {
                "name": "Dancing",
                "level": "begineer"
            }
        ]
    },
    {
        "_id": "[email protected]",
        "age": 7,
        "hobbiles": [{
                "name": "Coding",
                "level": "begineer"
            },
            {
                "name": "Chess",
                "level": "begineer"
            }
        ]
    },
    {
        "_id": "[email protected]",
        "age": 20,
        "hobbiles": [{
                "name": "Singing",
                "level": "begineer"
            },
            {
                "name": "Dancing",
                "level": "begineer"
            }
        ]
    },
    {
        "_id": "[email protected]",
        "age": 21,
        "hobbiles": [{
                "name": "Coding",
                "level": "begineer"
            },
            {
                "name": "Dancing",
                "level": "Football"
            }
        ]
    }
]

 

Now I want to fetch documents where id IN ([email protected], [email protected]) and age is greater than 5. [operationally] hobiiles Football.

My expectations from output is I should get three documents: and if hobbies is not matching then also it should be fine but if hobbies matches then that document should be on top. Basically I want to match hobbies but its optional if it doesn't match then also I should get data based on prior clauses.

[[email protected], [email protected], [email protected]]

test3 on top because Football matches there, and test and test1 because age and id matches there.

CodePudding user response：

This requires using Should clause. Should is equivalent to "OR". So a document will be returned if it satisfies any one condition in should query.

For conditions on id and age I have used filter clause. It is equivalent to "AND" . Filter clause does not calculate score for matched documents so any document which matches "hobbiles.level" will be ranked higher.

Query

{
  "query": {
    "bool": {
      "minimum_should_match": 1, 
      "should": [
        {
          "term": {
            "hobbiles.level.keyword": {
              "value": "Football"
            }
          }
        },
        {
          "bool": {
            "filter": [
              {
                "terms": {
                  "id.keyword": [
                    "[email protected]",
                    "[email protected]"
                  ]
                }
              },
              {
                "range": {
                  "age": {
                    "gt": 5
                  }
                }
              }
            ]
          }
        }
      ]
    }
  }
}

Result

  "hits" : [
      {
        "_index" : "index8",
        "_type" : "_doc",
        "_id" : "qE06noMBfFiM6spcUTo4",
        "_score" : 1.3112575,
        "_source" : {
          "id" : "[email protected]",
          "age" : 21,
          "hobbiles" : [
            {
              "name" : "Coding",
              "level" : "begineer"
            },
            {
              "name" : "Dancing",
              "level" : "Football"
            }
          ]
        }
      },
      {
        "_index" : "index8",
        "_type" : "_doc",
        "_id" : "pE03noMBfFiM6spc4jr2",
        "_score" : 0.0,
        "_source" : {
          "id" : "[email protected]",
          "age" : 12,
          "hobbiles" : [
            {
              "name" : "Singing",
              "level" : "begineer"
            },
            {
              "name" : "Dancing",
              "level" : "begineer"
            }
          ]
        }
      },
      {
        "_index" : "index8",
        "_type" : "_doc",
        "_id" : "pU03noMBfFiM6spc6DqZ",
        "_score" : 0.0,
        "_source" : {
          "id" : "[email protected]",
          "age" : 7,
          "hobbiles" : [
            {
              "name" : "Coding",
              "level" : "begineer"
            },
            {
              "name" : "Chess",
              "level" : "begineer"
            }
          ]
        }
      }
    ]

CodePudding user response：

Tldr;

It can be achieved via bool queries.

Solution

PUT /_bulk
{"index":{"_index":"73935795", "_id":"[email protected]"}}
{"age":12,"hobbiles":[{"name":"Singing","level":"begineer"},{"name":"Dancing","level":"begineer"}]}
{"index":{"_index":"73935795", "_id":"[email protected]"}}
{"age":7,"hobbiles":[{"name":"Coding","level":"begineer"},{"name":"Chess","level":"begineer"}]}
{"index":{"_index":"73935795", "_id":"[email protected]"}}
{"age":20,"hobbiles":[{"name":"Singing","level":"begineer"},{"name":"Dancing","level":"begineer"}]}
{"index":{"_index":"73935795", "_id":"[email protected]"}}
{"age":21,"hobbiles":[{"name":"Coding","level":"begineer"},{"name":"Dancing","level":"Football"}]}

GET 73935795/_search
{
  "query": {
    "bool": {
      "filter": [
        {
          "range": {
            "age": {
              "gt": 5
            }
          }
        },
        {
          "terms": {
            "_id": [
              "[email protected]",
              "[email protected]"
            ]
          }
        }
      ],
      "should": [
        {
          "match": {
            "hobbiles.level": "football"
          }
        }
      ]
    }
  }
}