ElasticSearch: Aggregation over nested array

Question

I have some documents that look like this

     {
      "reg": "a",
      "timestamp": "2012-10-01T00:06:35+0000",
      "data": [
        {
          "d": 33.2,
          "name": "TEMP",
        },
        {
          "d": 129,
          "name": "PRESS",
        }
      ]
    }

Data is a nested array with 2 properties inside (name and d which is a double)

I want to aggregate the value (d) of TEMP to have buckets with mean values, for instance.

How can I do that? if I create a simple aggregation on data.d, the aggregation mixes PRESS and TEMP, that I do not want

Answer 1

Take a look at the nested datatype and the nested aggregation