I'd like to group documents by a field (cluster_id) and find the distribution of group sizes
Ex:
200 groups have size 1
150 have size 2
...
50 have size 30
I tried with
{
"size": 0,
"aggs" : {
"clusters": { "terms" : { "field": "selection.cluster_id", "size": 1000 } },
"stats_cluster": {
"stats_bucket": { "buckets_path": "clusters>_count" }
}
}
}
but selection.cluster_id cardinality is 3M, and don't need the huge bucket list in the result