I do a search in elastic search and get all the items, sorted by the "prGreater" field, for example. But I can have it multiple times in the query return.

Is it possible to search for all items but in this return get only 1 of each item?

This is my mapping:

{
    "mappings": {
        "properties": {
            "name": {
                "type": "text"
            },
            "nmId": {
                "type": "text"
            },
            "pcDay": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "vlPaperNeg": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "dhDay": {
                "type": "date",
                "format": "strict_date_time"
            },
            "prBegin": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "prGreater": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "prLowest": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "prUNegoci": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "prAvg": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "prClose": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "qtdCamp": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "qtdSell": {
                "type": "scaled_float",
                "scaling_factor": 1000000
            },
            "estdPaper": {
                "type": "text"
            }
        }
    }
}

Query:

{
    "size": 100,
    "query": {
        "bool": {
            "filter": [
                {
                    "range": {
                        "dhDay": {
                            "from": "2022-06-23T00:00:00.000 0000",
                            "to": "2022-06-24T13:33:04.432 0000",
                            "include_lower": true,
                            "include_upper": false,
                            "boost": 1.0
                        }
                    }
                }
            ],
            "adjust_pure_negative": true,
            "boost": 1.0
        }
    },
    "sort": [
        {
            "prGreater": {
                "order": "desc"
            }
        }
    ]
}

CodePudding user response：

You basically want to group the items, say by their "name" and then for each group, get the latest item, sorted in descending order by dhDay.

You need a nested top hits aggregation:

{
  "size": 0,
  "aggs": {
    "name": {
      "terms": {
        "field": "name",
        "size": 100
      },
      "aggs": {
        "dhday_max": {
          "top_hits": {
            "size": 1,
            "sort": [
              {
                  "dhday": {
                      "order": "desc"
                  }
              }
            ]
          }
        }
      }
    }
  }
}

THe subaggregation will contain the latest item of a given name, sorted on the dhDay. You can add more fields to the sort or use 'prgreater" or anything else that you may need.