CODEWITHSUNDEEP
searchelasticsearchnested
ColdFox
ColdFox

Reputation: 81

Querying nested objects

Just to let things clear, first day working with Elastic... Moving to the problem.

I started to create my index with 

curl -XPUT "http://localhost:9200/users" -d'
{
   "mappings": {
      "user": {
         "properties": {
            "education": {
               "type": "nested"
            },
            "job": {
               "type": "nested"
            }
         }
      }
   }
}'

and then 

curl -XPOST "http://localhost:9200/users/user/" -d'
   {
      "name": "User A",
      "education": [
         {
            "school": "School A1",
            "course": "Course A1"
         },
         {
            "school": "School A2",
            "course": "Course A2"
         }
      ]
   }'

The problem that I'm facing now is the query part. I'm trying to get results with:

curl -XPOST "http://localhost:9200/users/user/_search?pretty" -d'
{
   "query": {
      "filtered": {
         "query": {
            "match_all": {}
         },
         "filter": {
            "nested": {
               "path": "education",
               "filter": {
                  "bool": {
                     "must": [
                        {
                           "term": {
                              "education.school": "School A1"
                           }
                        }
                     ]
                  }
               }
            }
         }
      }
   }
}'

But nothing is getting returned.

Upvotes: 0

Views: 40

Answers (2)

asang
asang

Reputation: 51

Just leaving my 2 cents here. I would avoid using filtered query as it is being deprecated Check this in latest release of ES.

I'll just rewrite the above query without using filtered query

curl -XPOST "http://localhost:9200/users/user/_search?pretty" -d'
{
   "query": {
      "nested": {
         "path": "education",
         "query": {
            "bool": {
               "must": [
                  {
                     "query_string": {
                        "default_field": "education.school",
                        "query": "School A1",
                        "default_operator": "AND"
                     }
                  }
               ]
            }
         }
      }
   }
}'

I followed this doc to write above query.

Upvotes: 0

Richa
Richa

Reputation: 7649

As per the mappings provided by you, school field is analyzed. Analyzed means the text School A will split over space and will be tokenized as School and A. you are searching using term query which looks for exact term. Study here about term query.

You can use Query_string with default_operator as AND

curl -XPOST "http://localhost:9200/users/user/_search?pretty" -d'
{
  "query": {
  "filtered": {
     "query": {
        "match_all": {}
     },
     "filter": {
        "nested": {
           "path": "education",
           "filter": {
              "bool": {
                 "must": [
                    {
                       "query": {
                          "query_string": {
                             "default_field": "education.school",
                             "query": "School A1",
                             "default_operator": "AND"
                          }
                       }
                    }
                 ]
              }
             }
        }
     }
     }
    }
 }'

Upvotes: 1

Related Questions