Elasticsearch aggregation only on specific entries in an array - arrays

Question

I'm new to Elasticsearch and can't figure out how to solve the following problem.
The easiest way to explain my problem is to show you an example.
The following array "listing" is part of all my files in Elasticsearch, but the entries vary, so the "person" with the "id" 42, might be in 50% of my files. What I'm trying to do is to get the average "ranking.position.standard" of all the persons with id 42 in all my files in Elasticsearch.
{
"listing": [
{
"person": {
"id": 42
},
"ranking": {
"position": {
"standard": 2
}
}
},
{
"person": {
"id": 55
},
"ranking": {
"position": {
"standard": 7
}
}
}
]
}
Thanks for your help!

Answer 1

First of all, do you store listings as an object or nested data type? I don't think it's going to work if it's an object, so try the following example:
PUT /test
{
"mappings": {
"_default_": {
"properties": {
"listing": {
"type": "nested"
}
}
}
}
}
PUT /test/aa/1
{
"listing": [
{
"person": {
"id": 42
},
"ranking": {
"position": {
"standard": 2
}
}
},
{
"person": {
"id": 55
},
"ranking": {
"position": {
"standard": 7
}
}
}
]
}
PUT /test/aa/2
{
"listing": [
{
"person": {
"id": 42
},
"ranking": {
"position": {
"standard": 5
}
}
},
{
"person": {
"id": 55
},
"ranking": {
"position": {
"standard": 6
}
}
}
]
}
GET test/_search
{
"size": 0,
"aggs": {
"nest": {
"nested": {
"path": "listing"
},
"aggs": {
"persons": {
"terms": {
"field": "listing.person.id",
"size": 10
},
"aggs": {
"avg_standard": {
"avg": {
"field": "listing.ranking.position.standard"
}
}
}
}
}
}
}
}
This has brought me the following result:
{
"took": 1,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 2,
"max_score": 0,
"hits": []
},
"aggregations": {
"nest": {
"doc_count": 4,
"persons": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": 42,
"doc_count": 2,
"avg_standard": {
"value": 3.5
}
},
{
"key": 55,
"doc_count": 2,
"avg_standard": {
"value": 6.5
}
}
]
}
}
}
}
It does seem correct.