Elastic Search Aggregate into buckets on conditions

Question

I am starting out with Elastic Search, and am stuck at trying to do some aggregation. Basically, I have a data set consisting of data in the following form:

{
    "name": "The Chef Restaurant",
    "city": "New York",
    "state": "New York",
    "rating": "GOOD",
    "type": "Continental"
}

Now, I want to do some aggregation and get all the Continental restaurants, Good restaurants, Restaurants in New York in one query.

Note that I don't want the count of all types of restaurants, I just want the count of the specific types. Also, these aggregations are mutually independent. That is, when I say GOOD, I don't necessarily want it to be Continental, it can be Italian or anything else.

This is what I have tried:

{
    "size": 0,
    "query": {
        "match_all": {}
    },
    "aggregations": {
        "good_restaurants": {
            "filters": {
                "match": {
                    "rating": "CONTINENTAL"
                }
            }
        },
        "continental_restaurants": {
            "filters": {
                "match": {
                    "type": "CONTINENTAL"
                }
            }
        },
        "restaurants_in_new_york": {
            "filters": {
                "match": {
                    "type": "CONTINENTAL"
                }
            }
        }
    }
}

which gives me the error:

{
   "error": {
      "root_cause": [
         {
            "type": "search_parse_exception",
            "reason": "Unknown key for a START_OBJECT in [good_restaurants]: [match].",
            "line": 9,
            "col": 17
         }
      ],
      "type": "search_phase_execution_exception",
      "reason": "all shards failed",
      "phase": "query",
      "grouped": true,
      "failed_shards": [
         {
            "shard": 0,
            "index": "test_master",
            "node": "-aWy78_mRaaBMcOAeiN9tg",
            "reason": {
               "type": "search_parse_exception",
               "reason": "Unknown key for a START_OBJECT in [good_restaurants]: [match].",
               "line": 9,
               "col": 17
            }
         }
      ]
   },
   "status": 400
}

I know this seems like a simple question, but I have been stuck at it for a long time. Any help will be appreciated.

Answer 1

You can make it work the way you expect, by doing it like this:

{
  "size": 0,
  "query": {
    "match_all": {}
  },
  "aggregations": {
    "selected_types": {
      "filters": {
        "filters": {
          "good_restaurants": {
            "match": {
              "rating": "CONTINENTAL"
            }
          },
          "continental_restaurants": {
            "match": {
              "type": "CONTINENTAL"
            }
          },
          "restaurants_in_new_york": {
            "match": {
              "type": "CONTINENTAL"
            }
          }
        }
      }
    }
  }
}