With Elasticsearch .NET and NEST 6.x : How to MultiGet documents from multiples indices

Question

var ids = new Dictionary<string, List<string>>();
ids["Topic"] = new List<string> {"KL7KJ2QBWD77yvpxyjvd", "374wJ2QBWD77yvpxDjpX", "B7499GMBWD77yvpxFzgW"};
ids["Prod"] = new List<string>();
ids["Account"] = new List<string>();

I made this fluent NEST query:

var results = client.MultiGet(m => 
m.Index("topics").GetMany<Topic>(ids["Topic"])

.Index("prods").GetMany<Prod>(ids["Prod"])

.Index("accounts").GetMany<Account>(ids["Account"])

it's generating the request below. We see the request use only the last index set which is "accounts" (and this is not what I need) :

http://localhost:9200/accounts/_mget?pretty=true
{
  "docs": [
    {
      "_type": "Topic",
      "_id": "KL7KJ2QBWD77yvpxyjvd"
    },
    {
      "_type": "Topic",
      "_id": "374wJ2QBWD77yvpxDjpX"
    },
    {
      "_type": "Topic",
      "_id": "B7499GMBWD77yvpxFzgW"
    }
  ]
}
# Response:
{
  "docs" : [
    {
      "_index" : "accounts",
      "_type" : "Topic",
      "_id" : "KL7KJ2QBWD77yvpxyjvd",
      "found" : false
    },
    {
      "_index" : "accounts",
      "_type" : "Topic",
      "_id" : "374wJ2QBWD77yvpxDjpX",
      "found" : false
    },
    {
      "_index" : "accounts",
          "_type" : "Topic",
      "_id" : "B7499GMBWD77yvpxFzgW",
      "found" : false
    }
  ]
}

In fact I would like to create in fluent NEST the following (valid) Elasticsearch query request (with no specific index) :

http://localhost:9200/_mget?pretty=true
{
  "docs": [
    {
      "_index": "topics",
      "_type": "Topic",
      "_id": "KL7KJ2QBWD77yvpxyjvd"
    },
    {
      "_index": "topics",
      "_type": "Topic",
      "_id": "374wJ2QBWD77yvpxDjpX"
    },
    {
      "_index": "topics",
      "_type": "Topic",
      "_id": "B7499GMBWD77yvpxFzgW"
    }
  ]
}

Is there a way, in fluent NEST, to specify each index/type for each list of ids ?

And of course if I'm adding specific ids to ids["Prod"] and ids["Account"] this will be generated properly for those... for instance:

http://localhost:9200/_mget?pretty=true
{
  "docs": [
    {
      "_index": "topics",
      "_type": "Topic",
      "_id": "KL7KJ2QBWD77yvpxyjvd"
    },
    {
      "_index": "prods",
      "_type": "Prod",
      "_id": "xxxxx"
    },
    {
      "_index": "accounts",
      "_type": "Account",
      "_id": "yyyyy"
    }
  ]
}

Answer 1

Each GetMany<T>(...) takes a second argument that is a delegate to further describe the call, including supplying an index name

var ids = new Dictionary<string, List<string>>
{
    { "Topic", new List<string> { "topic1", "topic2", "topic3" } },
    { "Prod", new List<string> { "prod1", "prod2", "prod3" } },
    { "Account", new List<string> { "account1", "account2", "account3" } }
};

var multiGetResponse = client.MultiGet(m => m
    .GetMany<Topic>(ids["Topic"], (op, id) => op
        .Index("topics")
    )
    .GetMany<Prod>(ids["Prod"], (op, id) => op
        .Index("prods")
    )
    .GetMany<Account>(ids["Account"], (op, id) => op
        .Index("accounts")
    )
);

which results in a request like

POST http://localhost:9200/_mget
{
  "docs": [
    {
      "_index": "topics",
      "_type": "topic",
      "_id": "topic1"
    },
    {
      "_index": "topics",
      "_type": "topic",
      "_id": "topic2"
    },
    {
      "_index": "topics",
      "_type": "topic",
      "_id": "topic3"
    },
    {
      "_index": "prods",
      "_type": "prod",
      "_id": "prod1"
    },
    {
      "_index": "prods",
      "_type": "prod",
      "_id": "prod2"
    },
    {
      "_index": "prods",
      "_type": "prod",
      "_id": "prod3"
    },
    {
      "_index": "accounts",
      "_type": "account",
      "_id": "account1"
    },
    {
      "_index": "accounts",
      "_type": "account",
      "_id": "account2"
    },
    {
      "_index": "accounts",
      "_type": "account",
      "_id": "account3"
    }
  ]
}

The "_type" names have been inferred from the CLR POCO names by lowercasing them, this can be changed by overriding the type name inferrer on Connection Settings.