String stats aggregation

edit

A multi-value metrics aggregation that computes statistics over string values extracted from the aggregated documents. These values can be retrieved either from specific keyword fields.

The string stats aggregation returns the following results:

  • count - The number of non-empty fields counted.
  • min_length - The length of the shortest term.
  • max_length - The length of the longest term.
  • avg_length - The average length computed over all terms.
  • entropy - The Shannon Entropy value computed over all terms collected by the aggregation. Shannon entropy quantifies the amount of information contained in the field. It is a very useful metric for measuring a wide range of properties of a data set, such as diversity, similarity, randomness etc.

For example:

response = client.search(
  index: 'my-index-000001',
  size: 0,
  body: {
    aggregations: {
      message_stats: {
        string_stats: {
          field: 'message.keyword'
        }
      }
    }
  }
)
puts response
POST /my-index-000001/_search?size=0
{
  "aggs": {
    "message_stats": { "string_stats": { "field": "message.keyword" } }
  }
}

The above aggregation computes the string statistics for the message field in all documents. The aggregation type is string_stats and the field parameter defines the field of the documents the stats will be computed on. The above will return the following:

{
  ...

  "aggregations": {
    "message_stats": {
      "count": 5,
      "min_length": 24,
      "max_length": 30,
      "avg_length": 28.8,
      "entropy": 3.94617750050791
    }
  }
}

The name of the aggregation (message_stats above) also serves as the key by which the aggregation result can be retrieved from the returned response.

Character distribution

edit

The computation of the Shannon Entropy value is based on the probability of each character appearing in all terms collected by the aggregation. To view the probability distribution for all characters, we can add the show_distribution (default: false) parameter.

response = client.search(
  index: 'my-index-000001',
  size: 0,
  body: {
    aggregations: {
      message_stats: {
        string_stats: {
          field: 'message.keyword',
          show_distribution: true
        }
      }
    }
  }
)
puts response
POST /my-index-000001/_search?size=0
{
  "aggs": {
    "message_stats": {
      "string_stats": {
        "field": "message.keyword",
        "show_distribution": true  
      }
    }
  }
}

Set the show_distribution parameter to true, so that probability distribution for all characters is returned in the results.

{
  ...

  "aggregations": {
    "message_stats": {
      "count": 5,
      "min_length": 24,
      "max_length": 30,
      "avg_length": 28.8,
      "entropy": 3.94617750050791,
      "distribution": {
        " ": 0.1527777777777778,
        "e": 0.14583333333333334,
        "s": 0.09722222222222222,
        "m": 0.08333333333333333,
        "t": 0.0763888888888889,
        "h": 0.0625,
        "a": 0.041666666666666664,
        "i": 0.041666666666666664,
        "r": 0.041666666666666664,
        "g": 0.034722222222222224,
        "n": 0.034722222222222224,
        "o": 0.034722222222222224,
        "u": 0.034722222222222224,
        "b": 0.027777777777777776,
        "w": 0.027777777777777776,
        "c": 0.013888888888888888,
        "E": 0.006944444444444444,
        "l": 0.006944444444444444,
        "1": 0.006944444444444444,
        "2": 0.006944444444444444,
        "3": 0.006944444444444444,
        "4": 0.006944444444444444,
        "y": 0.006944444444444444
      }
    }
  }
}

The distribution object shows the probability of each character appearing in all terms. The characters are sorted by descending probability.

Script

edit

If you need to get the string_stats for something more complex than a single field, run the aggregation on a runtime field.

POST /my-index-000001/_search
{
  "size": 0,
  "runtime_mappings": {
    "message_and_context": {
      "type": "keyword",
      "script": """
        emit(doc['message.keyword'].value + ' ' + doc['context.keyword'].value)
      """
    }
  },
  "aggs": {
    "message_stats": {
      "string_stats": { "field": "message_and_context" }
    }
  }
}

Missing value

edit

The missing parameter defines how documents that are missing a value should be treated. By default they will be ignored but it is also possible to treat them as if they had a value.

response = client.search(
  index: 'my-index-000001',
  size: 0,
  body: {
    aggregations: {
      message_stats: {
        string_stats: {
          field: 'message.keyword',
          missing: '[empty message]'
        }
      }
    }
  }
)
puts response
POST /my-index-000001/_search?size=0
{
  "aggs": {
    "message_stats": {
      "string_stats": {
        "field": "message.keyword",
        "missing": "[empty message]" 
      }
    }
  }
}

Documents without a value in the message field will be treated as documents that have the value [empty message].