how to store nested fields in elasticsearch

Question

I am using Java High Level REST client. Here is the link to its documentation

I have created a client.

trait HighLevelRestClient {

  def elasticSearchClient(): RestHighLevelClient = {
   new RestHighLevelClient(
      RestClient.builder(
        new HttpHost("localhost", ElasticSearchPort, "http")))
  }
}

While indexing the data, the nested fields are being stored as String. The following code explains how the index is being created:

val indexRequest = new IndexRequest("my-index", "test-type").source(
  "person", person,
  "date", DateTime.now()
)

where, person is a case class, represented as:

Person(personId: String, name: String, address: Address) 

and Address is itself a case class, represented as:

Address(city: String, zip: Int)

My application requires person to be stored as key-value pair, so that it's fields are searchable. But, when I am using the above code, it is being stored as String.

{
"person" : "Person(my-id, my-name, Address(my-city, zip-value))",
"date" :  "2017-12-12"
} 

and required structure is:

{
"person" : {
    "personId" : "my-id",
    "name" : "person-name",
    "address": {
      "city" : "city-name",
      "zip" : 12345
          }
     },
"date" : "2017-12-12"
}

I hope I have framed the question well. Any help would be appreciated. Thanks!

Answer 1

You are almost there. To achieve your goal you need to:

1. Serialize the object to JSON on your side
2. Specify the content type of the request

It is actually described in the page of the Index API.

A convenient library to serialize case classes into JSON is for example json4s (you can see some examples of serialization here).

Your code might look like the following:

import org.apache.http.HttpHost
import org.elasticsearch.action.index.IndexRequest
import org.elasticsearch.client.{RestClient, RestHighLevelClient}
import org.elasticsearch.common.xcontent.XContentType
import org.joda.time.DateTime
import org.json4s.NoTypeHints
import org.json4s.jackson.Serialization
import org.json4s.jackson.Serialization.write

case class Address(city: String, zip: Int)

case class Person(personId: String, name: String, address: Address)

case class Doc(person: Person, date: String)

object HighClient {
  def main(args: Array[String]): Unit = {
    val client = new RestHighLevelClient(
      RestClient.builder(
        new HttpHost("localhost", 9206, "http")))

    implicit val formats = Serialization.formats(NoTypeHints)

    val doc = Doc(
      Person("blah1", "Peter Parker", Address("New-York", 33755)),
      DateTime.now().toString
    )

    val indexRequest = new IndexRequest("my-index", "test-type").source(
      write(doc), XContentType.JSON
    )

    client.index(indexRequest)

    client.close()
  }
}

Note that in this case:

new IndexRequest("my-index", "test-type").source(
  write(doc), XContentType.JSON
)

this function will be used: public IndexRequest source(String source, XContentType xContentType)

While in your case:

new IndexRequest("my-index", "test-type").source(
  "person", person,
  "date", DateTime.now()
)

it will call public IndexRequest source(Object... source).

Hope that helps!