ElasticSearch 1x - aggregate on object conditions - elasticsearch

Question

I want to aggregate on data, which has inner objects. For example:
{
"_index": "product_index-en",
"_type": "elasticproductmodel",
"_id": "000001111",
"_score": 6.3316255,
"_source": {
"productId": "11111111111",
"productIdOnlyLetterAndDigit": "11111111111",
"productIdOnlyDigit": "11111111111",
"productNumber": "11111111111",
"name": "Glow Plug",
"nameOnlyLetterAndDigit": "glowplug",
"productImageLarge": "11111111111.jpg",
"itemGroupId": "11111",
"relatedProductIds": [],
"dataAreaCountries": [
"fra",
"pol",
"uk",
"sie",
"sve",
"atl",
"ita",
"hol",
"dk"
],
"oemItems": [
{
"manufactorName": "BERU",
"manufacType": "0"
},
{
"manufactorName": "LUCAS",
"manufacType": "0"
}
]
}
}
I need to be able aggregates oemItems.manufactorName values, but only where oemItems.manufacType is "0". I have tried a number of examples, such as the accepted one here ( Elastic Search Aggregate into buckets on conditions ), but I just cannot seem to wrap my head around it.
I've tried following, hopeing it will aggragate on manufacType first, which it does, and then manufactorName for each type, which it seems to display correct hit count. However, buckets for manufactorName are empty:
GET /product_index-en/_search
{
"size": 0,
"aggs": {
"baked_goods": {
"nested": {
"path": "oemItems"
},
"aggs": {
"test1": {
"terms": {
"field": "oemItems.manufacType",
"size": 500
},
"aggs": {
"test2": {
"terms": {
"field": "oemItems.manufactorName",
"size": 500
}
}
}
}
}
}
}
}
And the result:
{
"took": 27,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 471214,
"max_score": 0,
"hits": []
},
"aggregations": {
"baked_goods": {
"doc_count": 677246,
"test1": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "0",
"doc_count": 436557,
"test2": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": []
}
},
{
"key": "1",
"doc_count": 240689,
"test2": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": []
}
}
]
}
}
}
}
I have also tried to add a nested term filter, to only look at oemItems which have manufacType 1 with following query. However, it returns Objects where oemItems include manufacType 1, meaning it oemItems within products still contain either 1 or 0 manufacType. I don't see how doing an aggregate on this response will only return oemItems.manufactorName where oemItems.manufacType is 0
GET /product_index-en/_search
{
"query" : { "match_all" : {} },
"filter" : {
"nested" : {
"path" : "oemItems",
"filter" : {
"bool" : {
"must" : [
{
"term" : {"oemItems.manufacType" : "1"}
}
]
}
}
}
}
}

Answer 1

Good start so far. Just try it like this:
POST /product_index-en/_search
{
"size": 0,
"query": {
"nested": {
"path": "oemItems",
"query": {
"term": {
"oemItems.manufacType": "0"
}
}
}
},
"aggs": {
"baked_goods": {
"nested": {
"path": "oemItems"
},
"aggs": {
"test1": {
"terms": {
"field": "oemItems.manufactorName",
"size": 500
}
}
}
}
}
}