搜索排序通过它查询在弹性搜索中对单个文档的_source内的数据进行分页

Question

In an Elastic Search index we have a multiple document but one document is different from others. It contains vendor product mapping which no other document has, so it is one of its kind in the index. Like this:

    {
      "_index": "portal_support_20200911",
      "_type": "_doc",
      "_id": "techno_products",
      "_version": 20220829,
      "_seq_no": 39,
      "_primary_term": 1,
      "found": true,
      "_source": {
        "doc_name": "techno_products",
        "updated_on": "20220829",
        "products": [
          {
            "vendor": "Apple",
            "product": "Iphone"
          },
          {
            "vendor": "Samsung",
            "product": "Galaxy Z"
          },
          {
            "vendor": "Volkswagen",
            "product": "Passat"
          },
          {
            "vendor": "Volkswagen",
            "product": "Tiguan"
          }
        ]
      }
    }

It has thousands of vendor product mapping in "products" array. Our requirement is to write an Elastic Search query to get the data of this document but we also want to search, sort and paginate elements of this "products" array through ES query. Means we want to search, sort and paginate _source of this ES Document through ES query.

For example sort the "products" array by "vendor" or by "product" in desc order. Since this array has thousands of entries so we want to paginate and get 100 elements at a time then in next page next 100 and so on. We also want to give search option like vendor=Volkswagen, so it gives only matching elements in output from Elastic Search query.

I am new to ES but as per my knowledge we can search, sort and paginate documents in ES by their fields.

But can we also search, sort and paginate data inside _source of a document in Elastic Search? How can I achieve this using Elastic search query?

Please help me in this thanks.

Answer 1

I think you're looking for inner hits ?

If you have a nested query, you can add inner_hits as option:

{
  nested: {
    path: "products",
    query: {
      term: {
        "field": "term"
      }
    },
    inner_hits: {
      size: 10000,
      sort: [
        {
          "products.product": "desc",
        },
      ]
    }
  }
}

If you never use a nested query to retrieve your documents, I think you can add the nested query anyway and use match_all: {} inside the nested query:

{
  nested: {
    path: "products",
    query: {
      match_all: {}
    },
    inner_hits: {
      size: 10000,
      sort: [
        {
          "products.product": "desc",
        },
      ]
    }
  }
}

You can use from in inner_hits to page through the results.

I hope this helps.