Count emails that have more than 10 documents (buckets count)

[jack_daniels00]

Hi!
I have documents in elastic like:

{
    "email": "user@example.com",
    "subject": "Email subject",
    "body": "Email body"
}

And I'm trying to count all unique emails that have more than 10 documents.
I can retrieve all such emails by the aggregation query:

GET emails/_search
{
  "size": 0,
  "aggs": {
    "email_with_more_than_10_docs": {
      "terms": {
        "field": "email",
        "min_doc_count": 10
      }
    }
  }
}

And I get the output:

{
  "took" : 3,
  "timed_out" : false,
  "_shards" : {
    "total" : 3,
    "successful" : 3,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 74,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "aggregations" : {
    "email_with_more_than_10_docs" : {
      "doc_count_error_upper_bound" : 0,
      "sum_other_doc_count" : 0,
      "buckets" : [
        {
          "key" : "user1@example.com",
          "doc_count" : 40
        },
        {
          "key" : "user2@example.com",
          "doc_count" : 14
        },
        {
          "key" : "user3@example.com",
          "doc_count" : 13
        }
      ]
    }
  }
}

But how can I get the number of the unique emails (number of buckets)? In the example of output above that number is 3.

[Tomo_M (Tomohiro Mitani)]

As long as the size parameter of aggregation is larger than the number of emails which meet the criteria, count the length of the buckets array on the client side is a viable option to implement.

If you really need the count in the response, you may use stats_bucket pipeline aggregation. Keep in mind set "size" larger than the number of terms. The "count" value is what you want.

GET kibana_sample_data_flights/_search?filter_path=aggregations.stats
{
  "size":0,
  "aggs": {
    "dest": {
      "terms": {
        "field": "DestAirportID",
        "size": 10
      }
    },
    "stats":{
      "stats_bucket": {
        "buckets_path": "dest>_count"
      }
    }
  }
}

{
  "aggregations" : {
    "stats" : {
      "count" : 10,
      "min" : 305.0,
      "max" : 691.0,
      "avg" : 416.1,
      "sum" : 4161.0
    }
  }
}

GET kibana_sample_data_flights/_search?filter_path=aggregations.stats
{
  "size":0,
  "aggs": {
    "dest": {
      "terms": {
        "field": "DestAirportID",
        "size": 1000
      }
    },
    "stats":{
      "stats_bucket": {
        "buckets_path": "dest>_count"
      }
    }
  }
}

{
  "aggregations" : {
    "stats" : {
      "count" : 156,
      "min" : 1.0,
      "max" : 691.0,
      "avg" : 83.71153846153847,
      "sum" : 13059.0
    }
  }
}

[jack_daniels00]

But what if there are a lot of documents in elstic? Is there any other solution that takes it into account?

[Tomo_M (Tomohiro Mitani)]

For pagination of aggregation results, Composite aggregation may help you.

Using transform to perform aggregation beforehand might be another option.