我有以下索引映射：
{
"test5" : {
"mappings" : {
"dynamic" : "false",
"properties" : {
"messageType" : {
"type" : "keyword"
},
"groupId" : {
"type" : "keyword"
},
"payload" : {
"type" : "nested",
"include_in_root" : true,
"properties" : {
"request" : {
"type" : "nested",
"include_in_root" : true,
"properties" : {
"data" : {
"type" : "nested",
"include_in_root" : true,
"properties" : {
"chargingPeriods" : {
"type" : "nested",
"include_in_root" : true,
"properties" : {
"endDateTime" : {
"type" : "date"
},
"power" : {
"type" : "double"
},
"startDateTime" : {
"type" : "date"
}
}
}
}
}
}
}
}
}
}
}
}
}
第一個用例，我希望基于 2 分鐘間隔的存盤桶，payload.request.data.chargingPeriods.startDateTime過濾groupId 條件為messageType . BTWchargingPeriods 是一個陣列。
此查詢適用于該用例：
GET test5/_search
{
"size": 0,
"aggs": {
"my_buckets": {
"composite": {
"sources": [
{ "sessionId": { "terms": { "field": "groupId"} } },
{
"date" : {
"date_histogram": {
"field": "payload.request.data.chargingPeriods.startDateTime",
"fixed_interval": "2m",
"format": "MM/dd/yyyy - hh:mm:ss",
"order": "asc"
}
}
}
]
}
}
},
"query": {
"terms": {
"messageType": [
"test"
]
}
}
}
現在我希望在回傳的這些復合存盤桶上完成度量聚合，我嘗試了這個：
GET test5/_search
{
"size": 0,
"aggs": {
"my_buckets": {
"composite": {
"sources": [
{ "sessionId": { "terms": { "field": "groupId"} } },
{
"date" : {
"date_histogram": {
"field": "payload.request.data.chargingPeriods.startDateTime",
"fixed_interval": "2m",
"format": "MM/dd/yyyy - hh:mm:ss",
"order": "asc"
}
}
}
]
},
"aggregations": {
"metricAgg": {
"max": {
"field": "payload.request.data.chargingPeriods.power"
}
}
}
}
},
"query": {
"terms": {
"messageType": [
"test"
]
}
}
}
根據 ES 檔案https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-bucket-composite-aggregation.html，這應該通過在復合存盤桶上進行度量聚合來作業，但不是在復合桶上計算的度量聚合，它是在整個給定檔案中的chargingPeriods 陣列中的所有功率欄位中計算的。
我是如何創建索引的：
PUT /test5
{
"settings": {
"number_of_shards": 1
},
"mappings" : {
"dynamic" : "false",
"properties" : {
"groupId" : {
"type" : "keyword"
},
"messageType" : {
"type" : "keyword"
},
"payload" : {
"type" : "nested",
"include_in_root": true,
"properties": {
"request": {
"type":"nested",
"include_in_root":true,
"properties": {
"data": {
"type":"nested",
"include_in_root": true,
"properties": {
"chargingPeriods": {
"type": "nested",
"include_in_root": true,
"properties" : {
"endDateTime":{
"type": "date"
},
"power": {
"type": "double"
},
"startDateTime":{
"type": "date"
}
}
}
}
}
}
}
}
}
}
}
}
測驗資料：
POST test5/_doc/testdocu1
{
"groupId": "563",
"messageType": "test",
"payload": {
"request": {
"data": {
"chargingPeriods": [
{
"endDateTime": "2022-10-13T17:42:25Z",
"power": 9.62857,
"startDateTime": "2022-10-13T17:41:55Z"
},
{
"endDateTime": "2022-10-13T17:42:55Z",
"power": 9.6491,
"startDateTime": "2022-10-13T17:42:25Z"
},
{
"endDateTime": "2022-10-13T17:43:25Z",
"power": 9.6491,
"startDateTime": "2022-10-13T17:42:55Z"
},
{
"endDateTime": "2022-10-13T17:43:55Z",
"power": 9.66963,
"startDateTime": "2022-10-13T17:43:25Z"
},
{
"endDateTime": "2022-10-13T17:44:25Z",
"power": 9.67128,
"startDateTime": "2022-10-13T17:43:55Z"
},
{
"endDateTime": "2022-10-13T17:44:55Z",
"power": 9.65079,
"startDateTime": "2022-10-13T17:44:25Z"
},
{
"endDateTime": "2022-10-13T17:45:25Z",
"power": 9.66492,
"startDateTime": "2022-10-13T17:44:55Z"
},
{
"endDateTime": "2022-10-13T17:45:55Z",
"power": 9.68544,
"startDateTime": "2022-10-13T17:45:25Z"
},
{
"endDateTime": "2022-10-13T17:46:25Z",
"power": 9.68544,
"startDateTime": "2022-10-13T17:45:55Z"
},
{
"endDateTime": "2022-10-13T17:46:55Z",
"power": 9.67434,
"startDateTime": "2022-10-13T17:46:25Z"
}
]
}
}
}
}
我的輸出：
"aggregations" : {
"my_buckets" : {
"after_key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:46:00"
},
"buckets" : [
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:40:00"
},
"doc_count" : 1,
"metricAgg" : {
"value" : 9.68544
}
},
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:42:00"
},
"doc_count" : 4,
"metricAgg" : {
"value" : 9.68544
}
},
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:44:00"
},
"doc_count" : 4,
"metricAgg" : {
"value" : 9.68544
}
},
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:46:00"
},
"doc_count" : 1,
"metricAgg" : {
"value" : 9.68544
}
}
]
}
}
如您所見，它從所有元素中選擇了最大 payload.request.data.chargingPeriods.power，忽略了復合存盤桶。例如
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:40:00"
},
"doc_count" : 1,
"metricAgg" : {
"value" : 9.68544
}
},
metricAgg 應該是 9.62857
uj5u.com熱心網友回復：
它不能按您期望的方式作業，因為您正在聚合您擁有include_in_root的嵌套資料，因此，所有嵌套資料都在根檔案中找到，就好像它沒有嵌套一樣，因此，thestartDateTime和 the之間的關系power是基本輸了。
另一個問題是您的復合聚合聚合了嵌套 ( payload...) 和非嵌套資料 ( groupId)，這是行不通的。
但是，如果您在groupId陣列的每個元素中添加欄位，則可以使查詢作業如下：
GET test5/_search
{
"size": 0,
"aggs": {
"payload": {
"nested": {
"path": "payload"
},
"aggs": {
"request": {
"nested": {
"path": "payload.request"
},
"aggs": {
"data": {
"nested": {
"path": "payload.request.data"
},
"aggs": {
"charging": {
"nested": {
"path": "payload.request.data.chargingPeriods"
},
"aggs": {
"my_buckets": {
"composite": {
"sources": [
{
"sessionId": {
"terms": {
"field": "payload.request.data.chargingPeriods.groupId"
}
}
},
{
"date": {
"date_histogram": {
"field": "payload.request.data.chargingPeriods.startDateTime",
"fixed_interval": "2m",
"format": "MM/dd/yyyy - hh:mm:ss",
"order": "asc"
}
}
}
]
},
"aggregations": {
"metricAgg": {
"max": {
"field": "payload.request.data.chargingPeriods.power"
}
}
}
}
}
}
}
}
}
}
}
}
},
"query": {
"terms": {
"messageType": [
"test"
]
}
}
}
結果：
{
"key" : {
"sessionId" : "563",
"date" : "10/13/2022 - 05:40:00"
},
"doc_count" : 1,
"metricAgg" : {
"value" : 9.62857
}
},
轉載請註明出處，本文鏈接：https://www.uj5u.com/houduan/526735.html
標籤：弹性搜索弹性搜索聚合