Elasticsearch, get average document length

Question

Is there any better way in elasticsearch (other than issuing a match all query and manually averaging over the length of all returned documents) to get the average document

Answer 1

In ElasticSearch 6.2 you should just use the following line (no need to add 'terms'):

  "aggs" : 
      {"avg_size" : 
           {"avg" : 
               {"field" : "_size"}}}

See details here: https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-metrics-avg-aggregation.html

Answer 2

Shot in the dark, but facets or aggregations combined with a script might do it.

{
    ...,

    "aggs" : {
        "avg_length" : { "avg" : { "script" : "doc['_all'].length" } }
    }
}

Answer 3

I have used this code (I have the _source enabled)

{
  "query" : {"match_all" : {}},
  "aggs":{
    "avg_length" : { "avg" : { "script" : "_source.toString().length()"}}
  }
}

Well, the chars .. .if the string are UTF-8 to get the bytes:

{
  "query" : {"match_all" : {}},
  "aggs":{
    "avg_length" : { "avg" : { "script" : "_source.toString().getBytes(\"UTF-8\").length"}}
  }
}

Answer 4

The _size mapping field, if enabled, should give you the size of each document for free. Combining this with the avg aggregation should get you what you want. Something like:

{
  "query" : {"match_all" : {}},
  "aggs" : {"avg_size" : {"avg" : {"terms" : {"field" : "_size"}}}}
}