Fixing a Simple Elasticsearch Query

Question

I have below data:

{
   "results":[
      {
         "ID":"1",
         "products":[
            {
               "product":"car",
               "number":"5"
            },
            {
               "product":"computer",
               "number":"212"
            }
         ]
      },
      {
         "ID":"2",
         "products":[
            {
               "product":"car",
               "number":"9"
            },
            {
               "product":"computer",
               "number":"463"
            },
            {
               "product":"bicycle",
               "number":"5"
            }
         ]
      }
   ]
}

And my query is below:

{  
   "query":{  
      "bool":{  
         "must":[  
            {  
               "wildcard":{  
                  "results.products.product":"*car*"
               }
            },
            {  
               "wildcard":{  
                  "results.products.number":"*5*"
               }
            }
         ]
      }
   }
}

What I expect is to get only ID1. Because only it has a product with { "product":"car", "number":"5" } record. But what I get is both ID1 and ID2 because ID2's first record has "product":"car" and third record has "number":"5" records separately.

How can I fix this query?

Answer 1

You need to define your products as a nested type when creating mapping. Try with following mapping example:

PUT http://localhost:9200/indexname

{
  "mappings": {
     "typename": {
        "properties": {
           "products" : {
              "type" : "nested"
            }
          }
       }
    }
 }

Then you can use nested queries to match entire elements of your array - just as you need to.

{
  "query": {
    "nested": {
      "path": "products",
      "query": {
        "bool": {
          "must": [
            { "wildcard": { "products.product": "*car*" }},
            { "wildcard": { "products.number":  "*5*" }} 
          ]
        }
      }
    }
  }
}