I want to do an aggregation request with some additional logic, but I'm not sure if it's possible and how to do it. When I have the following documents, how can I requests the locations in the array that do not have a "type" defined? I tried with sending a sub aggregation filter on type, but then also location "20" get doc_count 1 on that aggregation.
How can I do some logic on the matching aggregation item in this case?
Documents:
//document1
{
"locations": [{
"code": "20",
"names": [{
"languageCode": "en-GB",
"value": "Amsterdam"
}
]
}, {
"type": {
"id": 25,
"names": [{
"languageCode": "en-GB",
"value": "area"
}
]
},
"code": "21",
"names": [{
"languageCode": "en-GB",
"value": "Amsterdam-South"
}
]
}
]
}
//Document 2
{
"locations": [{
"code": "22",
"names": [{
"languageCode": "en-GB",
"value": "DenHague"
}, {
"languageCode": "nl-NL",
"value": "DenHaag"
}
]
}
]
}
Request:
{
"aggs": {
"Filter_Location": {
"aggs": {
"SubType": {
"filter": {
"exists": {
"field": "locations.type"
}
}
} },
"terms": {
"field": "locations.code.keyword"
}
}
},
"size": 0
}
Result:
{
"aggregations": {
"Filter_Location": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "20",
"doc_count": 1,
"SubType": {
"doc_count": 1
},
"groupByAccoId": {
"value": 1
}
}, {
"key": "21",
"doc_count": 1,
"SubType": {
"doc_count": 1
},
"groupByAccoId": {
"value": 1
}
}, {
"key": "22",
"doc_count": 1,
"SubType": {
"doc_count": 0
},
"groupByAccoId": {
"value": 1
}
}
]
}
}
}
Expected Result:
{
"aggregations": {
"Filter_Location": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "20",
"doc_count": 1,
"SubType": {
"doc_count": 0
}
}, {
"key": "21",
"doc_count": 1,
"SubType": {
"doc_count": 1
}
}, {
"key": "22",
"doc_count": 1,
"SubType": {
"doc_count": 0
}
}
]
}
}
}
In order to prevent the locations
array flattening you'll need to set your index mapping as nested
:
PUT ind
{
"mappings": {
"properties": {
"locations": {
"type": "nested"
}
}
}
}
After ingesting the docs this query will fetch you the desired results:
GET ind/_search
{
"size": 0,
"aggs": {
"Filter_Location_parent": {
"nested": {
"path": "locations"
},
"aggs": {
"Filter_Location": {
"terms": {
"field": "locations.code.keyword"
},
"aggs": {
"SubType": {
"filter": {
"exists": {
"field": "locations.type"
}
}
}
}
}
}
}
}
}