Elasticsearch filter by nested fields

Question

I have a problem with creating a query to Elasticsearch with many conditions. My model looks like:

data class Product(
        @Id
        val id: String? = null,
        val category: String,
        val imagesUrls: List,
        @Field(type = FieldType.Double)
        val price: Double?,
        @Field(type = FieldType.Nested)
        val parameters: List?
)

data class Parameter(
        val key: String,
        val values: List
)

I would like to query products by:

category (for example cars)
price (between 20k $ and 50k $)
and parameters -> For example products with many parameters, like key capacity values 4L, 5L and second parameter gear transmission values manual

My current query looks like this:

GET data/_search
{

  "size": 10,
  "query": {
    "bool": {
      "must": [
        {
          "term": {
            "category.keyword": {
              "value": "cars"
            }
          }
        },
        {
          "nested": {
            "path": "parameters",
            "query": {
              "bool": {
                "must": [
                  {"term": {
                    "parameters.key.keyword": {
                      "value": "Capacity"
                    }
                  }},
                  {
                    "term": {
                      "parameters.key": {
                        "value": "4L, 5L"
                      }
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }

Could you tell me how to filter the product when parameter key is equal to Capacity and check that the values list contains one of the values?
How to combine many this kind operations in one query?

Example data:

{
   "category":"cars",
   "name":"Ferrari",
   "price":50000,
   "parameters":[
      {
         "key":"capacity",
         "values":"4L"
      },
      {
         "key":"gear transmission",
         "values":"automcatic"
      }
   ]
}

user_10348758 · Answer

The search query shown below queries the data based on:

category (for example cars)

And parameters -> For example products with many parameters, like key capacity values 4L, 5L and second parameter gear transmission values manual

Adding a working example with index data, mapping, search query, and search result

Index Mapping:

{
  "mappings": {
    "properties": {
      "parameters": {
        "type": "nested"
      }
    }
  }
}

Index Data:

{
   "category":"cars",
   "name":"Ferrari",
   "price":50000,
   "parameters":[
      {
         "key":"gear transmission",
         "values":["4L","5L"]
      },
      {
         "key":"capacity",
         "values":"automcatic"
      }
   ]
}

{
   "category":"cars",
   "name":"Ferrari",
   "price":50000,
   "parameters":[
      {
         "key":"capacity",
         "values":["4L","5L"]
      },
      {
         "key":"gear transmission",
         "values":"automcatic"
      }
   ]
}

{
   "category":"cars",
   "name":"Ferrari",
   "price":50000,
   "parameters":[
      {
         "key":"capacity",
         "values":"4L"
      },
      {
         "key":"gear transmission",
         "values":"automcatic"
      }
   ]
}

Search Query:

{
  "query": {
    "bool": {
      "must": [
        {
          "term": {
            "category.keyword": {
              "value": "cars"
            }
          }
        },
        {
          "nested": {
            "path": "parameters",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "parameters.key": "capacity"
                    }
                  },
                  {
                    "terms": {
                      "parameters.values": [
                        "4l",
                        "5l"
                      ]
                    }
                  }
                ]
              }
            }
          }
        },
        {
          "nested": {
            "path": "parameters",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "parameters.key": "gear transmission"
                    }
                  },
                  {
                    "match": {
                      "parameters.values": "automcatic"
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}

Search Result:

"hits": [
         {
            "_index": "bstof",
            "_type": "_doc",
            "_id": "1",
            "_score": 3.9281754,
            "_source": {
               "category": "cars",
               "name": "Ferrari",
               "price": 50000,
               "parameters": [
                  {
                     "key": "capacity",
                     "values": "4L"
                  },
                  {
                     "key": "gear transmission",
                     "values": "automcatic"
                  }
               ]
            }
         },
         {
            "_index": "bstof",
            "_type": "_doc",
            "_id": "2",
            "_score": 3.9281754,
            "_source": {
               "category": "cars",
               "name": "Ferrari",
               "price": 50000,
               "parameters": [
                  {
                     "key": "capacity",
                     "values": [
                        "4L",
                        "5L"
                     ]
                  },
                  {
                     "key": "gear transmission",
                     "values": "automcatic"
                  }
               ]
            }
         }
      ]

user_1136371 · Answer

When you need to match any one from a list then you can use terms query instead of term. Update the part in query from:

{
  "term": {
    "parameters.key": {
      "value": "4L, 5L"
    }
  }
}

to below:

{
  "terms": {
    "parameters.values": {
      "value": [
        "4L",
        "5L"
      ]
    }
  }
}

Note that if parameters.key is analysed field and there exist a keyword sub-field for the same, then use it instead. e.g parameters.values.keyword

You can read more on terms query here.

Elasticsearch filter by nested fields

2 Answers