ElasticSearch retrieve only list elements that match a term within single document

Question

Given this document structure.

{
   "title":"Lord of the rings",
   "id":"123abc",
   "pages":[
      {
         "pageNumber":1,
         "content":"Lorem ipsum lorem ipsum lorem ipsum"
      },
      {
         "pageNumber":2,
         "content":"dolor sit dolor sit dolor sit"
      }
   ]
}

and this query

{
   "query":{
      "match":{
         "pages.content":"lorem"
      }
   }
}

Is there a way that elastic can give me something like this as a result:

{
   "title":"Lord of the rings",
   "id":"123abc",
   "pages":[
      {
         "pageNumber":1,
         "content":"Lorem ipsum lorem ipsum lorem ipsum"
      }
   ]
}

So in case I have hundreds of pages I will only get those that match my term?

I though about creating a index for pages, but how would I get all the meta data from the document? Should this metadata be copied in all pages document from the same "book" lets say? What's the best approach in terms of performance?

Answer 1

Yup, documents sharing some common metadata is quite normal (and often necessary) in NoSQL datastores. As to an alternative approach, you could use highlighting or even some painless scripting but I'd recommend something along the lines of:

POST pages/_doc
{
  "book": {
    "title": "Lord of the rings",
    "id": "123abc",
    "metadata": {
      "num_of_pages": 300
    }
  },
  "page": {
    "pageNumber": 1,
    "content": "Lorem ipsum lorem ipsum lorem ipsum"
  }
}

POST pages/_doc
{
  "book": {
    "title": "Lord of the rings",
    "id": "123abc",
    "metadata": {
      "num_of_pages": 300
    }
  },
  "page": {
    "pageNumber": 2,
    "content": "dolor sit dolor sit dolor sit"
  }
}

And then

GET pages/_search
{
  "query": {
    "match": {
      "page.content": "lorem"
    }
  }
}

which is much simpler & scales reasonably well.