How variable width histogram with nested aggregations works

[rym (Mary)]

Hi,
I want to use the variable_width_histogram combined with other aggregations, such as min or max

Here is an example of combinated aggregations on a numeric field:

"aggs": {
        "aggregated-items": {
            "variable_width_histogram": {
                "buckets": 5,
                "field": "freespace"
            },
            "aggs": {
                "max-aggs": {
                    "max": {
                        "field": "freespace"
                    }
                }
            }
        }
    }

Here is the result I got:

 "aggregations": {
        "aggregated-items": {
          "buckets": [
                {
                    "min": 100.0,
                    "key": 111.67054263565892,
                    "max": 124.0,
                    "doc_count": 258,
                    "max-aggs": {
                        "value": 200.0
                    }
                },
                {
                    "min": 124.0,
                    "key": 134.45454545454547,
                    "max": 145.0,
                    "doc_count": 198,
                    "max-aggs": {
                        "value": 200.0
                    }
                },
                {
                    "min": 145.0,
                    "key": 157.73333333333332,
                    "max": 171.0,
                    "doc_count": 270,
                    "max-aggs": {
                        "value": 200.0
                    }
                },
                {
                    "min": 172.0,
                    "key": 180.82323232323233,
                    "max": 191.0,
                    "doc_count": 198,
                    "max-aggs": {
                        "value": 200.0
                    }
                },
                {
                    "min": 192.0,
                    "key": 196.11764705882354,
                    "max": 200.0,
                    "doc_count": 85,
                    "max-aggs": {
                        "value": 198.0
                    }
                }
            ]
        }
    }

I don't understant why the max-aggs value is 200.0 in first buckets and why it's 198.0 in the last bucket. I would expect the max-aggs to be the same as the max of the corresponding bucket, ie 124.0 in the first, etc.

So, I would like to know how this max-aggs value is calulcated when used under a variable_width_histogram ?

Thanks !

[Tomo_M (Tomohiro Mitani)]

Hi,

Does "freespace" field have multiple values in array?
As bucket aggregation works for bucketing documents, each variable_width_histogram bucket may contain documents with a value of "freespace" between the range. If the documents have other values for "freespace" field, the sub-aggregation uses all the values for aggregation.

If you have to aggregate such values separately, storing them in individual nested objects may be an solution.

[rym (Mary)]

Hi,

Thanks for your answer, now I understand better how it works when aggregating arrays. Infortunately, the "freespace" fields is not an array.

However, it works perfectly when I use an histogram. So I don't know why I don't have same max with variable_width_histogram.