问题
Here is my query -
{
"from": 0,
"size": 100,
"query": {
"bool": {
"filter": [
{
"terms": {
"folderId.keyword": [
"ff98505e-cdff-43aa-8b05-197bc3f3265e"
],
"boost": 1
}
},
{
"terms": {
"objectType.keyword": [
"File"
],
"boost": 1
}
},
{
"term": {
"tenantId": {
"value": "34202",
"boost": 1
}
}
}
],
"adjust_pure_negative": true,
"boost": 1
}
},
"aggs":{
"_byformat":{
"terms":{
"field":"format.keyword",
"min_doc_count":0,
"size":200
}
}
}
}
result with min_doc_count = 0 -
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 3,
"max_score" : 0.0,
"hits" : [
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "6adbda83-53ad-457f-a2ab-d5b04c643005",
"_score" : 0.0,
"_source" : {
"format" : "vnd.openxmlformats-officedocument.spreadsheetml.sheet",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New XLSX file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "6adbda83-53ad-457f-a2ab-d5b04c643005",
"status" : "active",
"expirationDate" : null
}
},
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "b1000a15-2d80-41f4-a5df-ba5c27f8e9c6",
"_score" : 0.0,
"_source" : {
"format" : "vnd.ms-excel",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New XLS file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "b1000a15-2d80-41f4-a5df-ba5c27f8e9c6",
"status" : "active",
"expirationDate" : null
}
},
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "630e9f49-3368-408d-a091-03f253127004",
"_score" : 0.0,
"_source" : {
"format" : "msword",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New DOC file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "630e9f49-3368-408d-a091-03f253127004",
"status" : "active",
"expirationDate" : null
}
}
]
},
"aggregations" : {
"_byformat" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "msword",
"doc_count" : 1
},
{
"key" : "vnd.ms-excel",
"doc_count" : 1
},
{
"key" : "vnd.openxmlformats-officedocument.spreadsheetml.sheet",
"doc_count" : 1
},
{
"key" : "bmp",
"doc_count" : 0
},
{
"key" : "gif",
"doc_count" : 0
},
{
"key" : "html",
"doc_count" : 0
}
]
}
}
}
result with min_doc_count = 1 -
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 3,
"max_score" : 0.0,
"hits" : [
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "6adbda83-53ad-457f-a2ab-d5b04c643005",
"_score" : 0.0,
"_source" : {
"format" : "vnd.openxmlformats-officedocument.spreadsheetml.sheet",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New XLSX file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "6adbda83-53ad-457f-a2ab-d5b04c643005",
"status" : "active",
"expirationDate" : null
}
},
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "b1000a15-2d80-41f4-a5df-ba5c27f8e9c6",
"_score" : 0.0,
"_source" : {
"format" : "vnd.ms-excel",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New XLS file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "b1000a15-2d80-41f4-a5df-ba5c27f8e9c6",
"status" : "active",
"expirationDate" : null
}
},
{
"_index" : "plnesdv1-34202-1",
"_type" : "_doc",
"_id" : "630e9f49-3368-408d-a091-03f253127004",
"_score" : 0.0,
"_source" : {
"format" : "msword",
"externalSharing" : "N",
"description" : null,
"dateModified" : null,
"type" : "application",
"folderId" : "ff98505e-cdff-43aa-8b05-197bc3f3265e",
"tags" : [ ],
"objectType" : "File",
"dateCreated" : null,
"name" : "New DOC file",
"tenantId" : "34202",
"modifiedBy" : "rdt001",
"id" : "630e9f49-3368-408d-a091-03f253127004",
"status" : "active",
"expirationDate" : null
}
}
]
},
"aggregations" : {
"_byformat" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "msword",
"doc_count" : 1
},
{
"key" : "vnd.ms-excel",
"doc_count" : 1
},
{
"key" : "vnd.openxmlformats-officedocument.spreadsheetml.sheet",
"doc_count" : 1
}
]
}
}
}
Aggregations are correct when min_doc_count = 1 and only buckets relevant to hits are fetched.
Could anyone tell me why aggregation is fetching all buckets when min_doc_count = 0 is set. I have gone through the elastic search documentation, it states that this behavior is by design, any other way to get aggregation buckets only for hits and with zero count as well.
回答1:
Seems you slightly misunderstand the min_doc_count on aggregation,
Setting min_doc_count=0 will also return buckets for terms that didn’t match any hit
It is possible to only return terms that match more than a configured number of hits using the min_doc_count option:
"aggs":{
"_byformat":{
"terms":{
"field":"format.keyword",
"min_doc_count":0,
"size":200
}
}
}
The above aggregation would only return keyword which has been found in 0 hits or more. 0 means all with/without, 1 means minimum 1 document should have that keyword.
回答2:
First of all you need to understand what is the meaning of seeing buckets with zero counts.
Below is an excerpt from the Terms Aggregation link:
Setting min_doc_count=0 will also return buckets for terms that didn’t match any hit. However, some of the returned terms which have a document count of zero might only belong to deleted documents or documents from other types, so there is no warranty that a match_all query would find a positive document count for those terms.
So most likely it appears to be the count for the deleted documents.
Note that the aggregation would only get calculated on the documents that get filtered by the query.
However you need to keep in mind that while ES keeps merging the segments of indexes behind the scenes(that happens during deletion process), the results for count with 0 may not be consistent and over a period of time may, eventually (if no further dos are deleted from that point onwards) you may not get any terms with 0 count at all once the merging process is completed.
So in a way it is safe to say to your business leads, that they are the counts for deleted docs and you can push the above argument to them. And if they say they need count of docs/terms of deleted docs, it is like finding a document/terms which does not exist in the index and it does not even make sense right.
As per why does this still shows, that is probably due to the segment merging process that happens in ES and it is by design.
So no, you cannot apply query/filter on deleted documents (take a step back and imagine that) and hence you cannot control the data related to docs not available in first place.
Aggregation Query:
You can make use of the below aggregation which would give you as per your requirement mentioned in the comment:
POST <your_index_name>/_search
{
"size": 0,
"aggs": {
"myaggs_count_zero": { <--- Agg for count 0
"terms": {
"field": "format.keyword"
},
"aggs": {
"document_counts": {
"value_count" : {
"field" : "format.keyword"
}
},
"by_account_filtered": {
"bucket_selector": {
"buckets_path": {
"totalDocs": "document_counts"
},
"script": "params.totalDocs == 0"
}
}
}
},
"myaggs_count_not_zero": { <--- Agg for normal count
"terms": {
"field": "format.keyword",
"min_doc_count": 1
}
}
}
}
Note that I've made use of Terms, Value Count and Bucket Selector Aggregations
This may not be what you are looking for but I hope that helps!
来源：https://stackoverflow.com/questions/62593947/elastic-search-aggregation-using-min-doc-count-0-returns-all-the-buckets-which-a