Hi,
I'm still quite new to elasticsearch and have found myself needing to use a nested index. I've tried to distill the problem into something simple that I could post here.
This comes from a database where each record has anywhere from 1 to a few hundred children. Each child has a unique name that comes from a pool of about 3000+ different names.
Mapping (from kibana console)
PUT test_nest
{
"mappings": {
"_doc": {
"dynamic": "false",
"properties": {
"full_name": {
"type": "text"
},
"v": {
"type": "nested",
"properties": {
"name": {
"type": "keyword"
},
"value": {
"type": "float"
}
}
}
}
}
}
}
Sample data
POST _bulk
{"index": {"_index": "test_nest", "_type": "_doc", "_id": "1"}}
{"full_name": "Three", "v": [{"name": "test_vol", "value": 3}, {"name": "duration", "value": 30}]}
{"index": {"_index": "test_nest", "_type": "_doc", "_id": "2"}}
{"full_name": "One", "v": [{"name": "test_vol", "value": 1}, {"name": "duration", "value": 10}]}
{"index": {"_index": "test_nest", "_type": "_doc", "_id": "3"}}
{"full_name": "Two", "v": [{"name": "test_vol", "value": 2}, {"name": "duration", "value": 20}]}
I'm trying to get an average value of all the v.name == test_vol. The query I use below does not work since it averages all the v.value entries.
I'm looking on how to limit the scope of the v.value entries to only those matching test_vol. Sorry if this is a newbie question but I'm at my wits end.
Aggregation Query
GET test_nest/_search
{
"size": 0,
"query": {
"nested": {
"path": "v",
"query": {
"bool": {
"must": [
{
"match": {
"v.name": "test_vol"
}
}
]
}
}
}
},
"aggs": {
"avg_test_vol": {
"nested": {
"path": "v"
},
"aggs": {
"thing": {
"avg": {
"field": "v.value"
}
}
}
}
}
}
Results of query. I'd want to get an average of 2. Instead I get 11.
{
"took": 3,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"skipped": 0,
"failed": 0
},
"hits": {
"total": 3,
"max_score": 0,
"hits": []
},
"aggregations": {
"avg_test_vol": {
"doc_count": 6,
"thing": {
"value": 11
}
}
}
}
Thanks in advance for any hints on how to scope the documents used for the aggregation,
Rick