elasticsearch - Elastic Search Sum aggregation with group by and where condition

Question

I am newbie in ElasticSearch.

We are currently moving our code from relational DB to ElasticSearch. So we are converting our queries in ElasticSearch query format.

I am looking for ElasticSearch equivalent of below query -

SELECT Color, SUM(ListPrice), SUM(StandardCost)
FROM Production.Product
WHERE Color IS NOT NULL 
    AND ListPrice != 0.00 
    AND Name LIKE 'Mountain%'
GROUP BY Color

Can someone provide me the example of ElasticSearch query for above?

Answer 1

You'd have a products index with a product type documents whose mapping could look like this based on your query above:

curl -XPUT localhost:9200/products -d '
{
  "mappings": {
    "product": {
      "properties": {
        "Color": {
          "type": "string"
        },
        "Name": {
          "type": "string"
        },
        "ListPrice": {
          "type": "double"
        },
        "StandardCost": {
          "type": "double"
        }
      }
    }
  }
}'

Then the ES query equivalent to the SQL one you gave above would look like this:

{
  "query": {
    "filtered": {
      "query": {
        "query_string": {
          "default_field": "Name",
          "query": "Mountain*"
        }
      },
      "filter": {
        "bool": {
          "must_not": [
            {
              "missing": {
                "field": "Color"
              }
            },
            {
              "term": {
                "ListPrice": 0
              }
            }
          ]
        }
      }
    }
  },
  "aggs": {
    "by_color": {
      "terms": {
        "field": "Color"
      },
      "aggs": {
        "total_price": {
          "sum": {
            "field": "ListPrice"
          }
        },
        "total_cost": {
          "sum": {
            "field": "StandardCost"
          }
        }
      }
    }
  }
}