在Elasticsearch中获取列表中对象的唯一结果

Question

我有这样的映射

"custom_metadata": {
"properties": {
"key": {
"type": "keyword"
},
"value": {
"type": "keyword"
}
}
}

摄入的数据看起来像这个

// data in document 1
"custom_metadata": [
{
"value": "NPL",
"key": "schema"
},
{
"value": "SAPERP",
"key": "system"
}
]

// data in document 2
"custom_metadata": [
{
"value": "trial",
"key": "schema"
},
{
"value": "Oracle",
"key": "system"
}
]

我想在每个关键字上进行聚合，并在搜索结果中获得相关值，比如这个

"schema": [
{"value": "NPL"},
{ "value": "trial",}
],
"system":[
{"value": "SAPERP"},
{ "value": "Oracle",}
]

注：以上输出仅供表示。如果我在ES中得到了类似的东西，那么我可以在服务端上解析并获得所需的结果

我尝试过的：

"custom_metadata_key": {
"terms": {
"field": "custom_metadata.key"
},
"aggregations": {
"custom_metadata_value": {
"terms": {
"field": "custom_metadata.value"
}
}
}
}

在嵌套的agg之上，聚合每个键并在结果中给出所有值。

{
"key" : "schema",
"doc_count" : 2,
"custom_metadata_value" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "Oracle",
"doc_count" : 2
},
{
"key" : "NPL",
"doc_count" : 1
},
{
"key" : "SAPERP",
"doc_count" : 1
},
{
"key" : "trial",
"doc_count" : 1
}
]
}
}

上面的输出为所有键重复，并为所有键提供相同的聚合。

Answer 1

您需要将字段custom_metadata的数据类型从object更改为nested，就可以轻松地实现您想要的输出。

映射

{
"mappings": {
"properties": {
"custom_metadata":{
"type": "nested"
}
}
}
}

查询

{
"size": 0,
"aggs": {
"data": {
"nested": {
"path": "custom_metadata"
},
"aggs": {
"custom_metadata_key": {
"terms": {
"field": "custom_metadata.key.keyword",
"size": 10
},
"aggs": {
"custom_metadata_value": {
"terms": {
"field": "custom_metadata.value.keyword",
"size": 10
}
}
}
}
}
}
}
}

输出

"aggregations": {
"data": {
"doc_count": 4,
"custom_metadata_key": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "schema",
"doc_count": 2,
"custom_metadata_value": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "NPL",
"doc_count": 1
},
{
"key": "trial",
"doc_count": 1
}
]
}
},
{
"key": "system",
"doc_count": 2,
"custom_metadata_value": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "Oracle",
"doc_count": 1
},
{
"key": "SAPERP",
"doc_count": 1
}
]
}
}
]
}
}
}