Combine multiple index and nested in search elasticsearch

Elastic Stack Elasticsearch
1

How to search multiple indexes with nested and non-nested fields ?
I have an audio index with the results like this :

{
        "_index" : "audios",
        "_type" : "_doc",
        "_id" : "145",
        "_score" : 9.143846,
        "_source" : {
	"id":145,
          "audio_title" : "Title",
          "main_artists" : [
            {
              "id" : 1,
              "artist_name" : "Artist_name"
            }
          ]
	}
}

and an index artists with results like this :

{
	  "_index" : "artists",
	  "_type" : "_doc",
	  "_id" : "34",
	  "_version" : 2,
	  "_seq_no" : 130,
	  "_primary_term" : 1,
	  "found" : true,
	  "_source" : {
		"id":34,
	   	 "artist_name" : "Artist_name"
	  }
}

My problem is when I do a search on the 2 indexes it gives me an error.

GET audios,artists/_search
{
  "query": {
    "bool": {
      "must": [
        {
          "bool": {
            "should": [
              {"match_phrase_prefix": {
                "audio_title": "XXXX"
              }},
              {"nested": {
                "path": "main_artists",
                "query": {
                  "match_phrase_prefix": {
                    "main_artists.artist_name": "XXXX"
                  }
                }
              }}
            ]
          }
        }
      ]
    }
  }
}

The error:

"failures" : [
      {
        "shard" : 0,
        "index" : "artists",
        "node" : "sHio0-yeRaO8Z5qJXaczPw",
        "reason" : {
          "type" : "query_shard_exception",
          "reason" : "failed to create query: [nested] failed to find nested object under path [main_artists]",
          "index_uuid" : "etDWonKCSQietG3uK3ZnBA",
          "index" : "artists",
          "caused_by" : {
            "type" : "illegal_state_exception",
            "reason" : "[nested] failed to find nested object under path [main_artists]"
          }
        }
      }
    ]
2

Hi @stephane_chan

You can create separate queries for each index and run in an msearch.