CODEWITHSUNDEEP
elasticsearch
Mohitt
Mohitt

Reputation: 2977

Elasticsearch DSL: Aggregation

Shown below is the structure of type (kind of a table) of my data.

 Aircraft | Duration 
 A320     | 0.95
 A320     | 0.55
 A321     | 16.50
 A321     | 3.9

In this data, I want to perform a ceil() on duration, followed by a groupBy operation to get following output:

 Aircraft | Duration | Count
 A320     | 1        |  2
 A321     | 17       |  1
 A321     | 4        |  1

Upvotes: 0

Views: 443

Answers (1)

Val
Val

Reputation: 217254

I've based myself on the following mapping type (edited Duration as string):

curl -XPUT localhost:9200/tests -d '
{
  "mappings": {
    "test1": {
      "properties": {
        "Aircraft": {
          "type": "string"
        },
        "Duration": {
          "type": "string"
        }
      }
    }
  }
}'

And created four documents that match your data table above.

curl -XPOST localhost:9200/tests/_bulk -d '
{"index": {"_type": "test1", "_id": 1}}
{"Aircraft": "A320", "Duration": "0.95"}
{"index": {"_type": "test1", "_id": 2}}
{"Aircraft": "A320", "Duration": "0.55"}
{"index": {"_type": "test1", "_id": 3}}
{"Aircraft": "A321", "Duration": "16.50"}
{"index": {"_type": "test1", "_id": 4}}
{"Aircraft": "A321", "Duration": "3.9"}
'

The aggregation query that will return the results you expect would look like this:

curl -XPOST localhost:9200/tests/_search -d '
{
  "size": 0,
  "query": {
    "filtered": {
      "filter": {
        "terms": {
          "Aircraft": [
            "a320",
            "b737"
          ]
        }
      }
    }
  },
  "aggs": {
    "aircrafts": {
      "terms": {
        "field": "Aircraft"
      },
      "aggs": {
        "duration": {
          "terms": {
            "script": "Math.ceil(doc['Duration'].value as double)"
          }
        }
      }
    }
  }
}'

The output of that query looks like this:

enter image description here

Note: Make sure to enable scripting in your elasticsearch.yml file by adding

script.disable_dynamic: false

Upvotes: 1

Related Questions