Elasticsearch DSL 查询 - 可选的术语和分数
Elasticsearch DSL queries - optional should terms & scores
我是 Elasticsearch 世界的新手，我可能遗漏了一些概念。
这是我不理解的场景：
我想根据以下条件查找文档：
- category.level = A
- category.name =“John.G”或“Chris T.”
- 批准 = 是（可选）
映射：
PUT data
{
"mappings": {
"properties": {
"createdAt": {
"type": "date",
"format": "yyyy-MM-dd HH:mm:ss.SSSZ"
},
"category": {
"type": "nested",
"properties": {
"name": {
"type": "text",
"analyzer": "keyword"
}
}
},
"approved": {
"type": "text",
"analyzer": "keyword"
}
}
}
}
数据：
POST data/_create/1
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Mary F.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "yes"
}
POST data/_create/2
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
POST data/_create/3
{
"category": [
{
"name": "John G.",
"level": "C"
},
{
"name": "Phil C.",
"level": "C"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
POST data/_create/4
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2020-04-18 19:09:27.527+0200",
"approved": "yes"
}
POST data/_create/5
{
"category": [
{
"name": "Unknown A.",
"level": "A"
},
{
"name": "Unknown B.",
"level": "A"
}
],
"createdBy": "Unknown",
"createdAt": "2020-08-18 19:09:27.527+0200",
"approved": "yes"
}
查询：
GET data/_search
{
"query": {
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{"match": {"category.level": "A"}}
],
"should": [
{"term": {"category.name": "John G."}},
{"term": {"category.name": "Chris T."}},
{"term": {"approved": "yes"}}
],
"minimum_should_match": 1
}
}
}
}
}
回复：
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 3,
"relation" : "eq"
},
"max_score" : 1.4455402,
"hits" : [
{
"_index" : "data",
"_id" : "2",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "no"
}
},
{
"_index" : "data",
"_id" : "4",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2020-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_id" : "1",
"_score" : 1.151647,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Mary F.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "yes"
}
}
]
}
}
问题：
- 为什么返回的第一个文档是
approval = no？我期待带有 approval = yes 的文档会得到更好的评分。
- 为什么索引 = 5 的文档（它不符合条件
category.name，但符合 approved = yes）没有被返回？
approved = yes 的可选性未在上述查询中表达。我怎样才能用 minimum_should_match: 0 创建一种额外的分隔 should 术语？可以增加分数但不会过滤结果的东西。
我用你的映射和示例数据尝试了你的场景，发现了问题，你在导致问题的 nested 查询上下文中使用了 approved:yes，这就是导致问题的原因，如果您将查询更改为下面的 （基本上在 should 块中使用 approved:yes 但在 nested 查询之外），它会解决您所有的问题。
{
"query": {
"bool": {
"should": [
{
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{
"match": {
"category.level": "A"
}
}
],
"should": [
{
"term": {
"category.name": "John G."
}
},
{
"term": {
"category.name": "Chris T."
}
}
]
}
}
}
},
{
"term": {
"approved": "yes"
}
}
]
}
}
}
和搜索结果
"hits": [
{
"_index": "71967271",
"_id": "4",
"_score": 1.9845366,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2020-04-18 19:09:27.527+0200",
"approved": "yes"
}
},
{
"_index": "71967271",
"_id": "2",
"_score": 1.4455402,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
},
{
"_index": "71967271",
"_id": "1",
"_score": 1.2437345,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Mary F.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "yes"
}
},
{
"_index": "71967271",
"_id": "5",
"_score": 0.7968255,
"_source": {
"category": [
{
"name": "Unknown A.",
"level": "A"
},
{
"name": "Unknown B.",
"level": "A"
}
],
"createdBy": "Unknown",
"createdAt": "2020-08-18 19:09:27.527+0200",
"approved": "yes"
}
}
]
您需要使用下面的查询，其中有主要的 bool 查询。它有第一个带有嵌套查询的 must 子句，它有 bool 查询 category.level 字段，然后另一个 bool 查询带有 should 子句 category.name字段。
现在主 bool 查询有 approved 的 should 子句，用于提升具有 yes 值的结果（这在 nested 查询之外）。
POST data/_search
{
"query": {
"bool": {
"must": [
{
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{
"term": {
"category.level": {
"value": "a"
}
}
},
{
"bool": {
"should": [
{
"term": {
"category.name": "John G."
}
},
{
"term": {
"category.name": "Chris T."
}
}
]
}
}
]
}
}
}
}
],
"should": [
{
"term": {
"approved": "yes"
}
}
]
}
}
}
结果：
{
"took" : 2,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 3,
"relation" : "eq"
},
"max_score" : 1.9845366,
"hits" : [
{
"_index" : "data",
"_type" : "_doc",
"_id" : "4",
"_score" : 1.9845366,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2020-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.6906434,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Mary F.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_type" : "_doc",
"_id" : "2",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "no"
}
}
]
}
}
Why the first document returned is an approval = no? I was expecting
that docs with approval = yes would be better scored.
因为您在 nested 查询中有 should 子句，并且它与任何文档都不匹配，因为 approved 在 category 之外，因此它不会改变分数。
Why doc with index = 5 (it doesn't attend the criteria category.name,
but it does for approved = yes) is not being returned?
它已被您的 must 子句删除，但是如果您还需要 index =5 文档，那么您可以添加两个 should 子句，一个用于嵌套，一个用于 approved，它将解决你的问题。
我的回答也解决了你的问题3。
我是 Elasticsearch 世界的新手，我可能遗漏了一些概念。
这是我不理解的场景：
我想根据以下条件查找文档：
- category.level = A
- category.name =“John.G”或“Chris T.”
- 批准 = 是（可选）
映射：
PUT data
{
"mappings": {
"properties": {
"createdAt": {
"type": "date",
"format": "yyyy-MM-dd HH:mm:ss.SSSZ"
},
"category": {
"type": "nested",
"properties": {
"name": {
"type": "text",
"analyzer": "keyword"
}
}
},
"approved": {
"type": "text",
"analyzer": "keyword"
}
}
}
}
数据：
POST data/_create/1
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Mary F.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "yes"
}
POST data/_create/2
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
POST data/_create/3
{
"category": [
{
"name": "John G.",
"level": "C"
},
{
"name": "Phil C.",
"level": "C"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
POST data/_create/4
{
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2020-04-18 19:09:27.527+0200",
"approved": "yes"
}
POST data/_create/5
{
"category": [
{
"name": "Unknown A.",
"level": "A"
},
{
"name": "Unknown B.",
"level": "A"
}
],
"createdBy": "Unknown",
"createdAt": "2020-08-18 19:09:27.527+0200",
"approved": "yes"
}
查询：
GET data/_search
{
"query": {
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{"match": {"category.level": "A"}}
],
"should": [
{"term": {"category.name": "John G."}},
{"term": {"category.name": "Chris T."}},
{"term": {"approved": "yes"}}
],
"minimum_should_match": 1
}
}
}
}
}
回复：
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 3,
"relation" : "eq"
},
"max_score" : 1.4455402,
"hits" : [
{
"_index" : "data",
"_id" : "2",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "no"
}
},
{
"_index" : "data",
"_id" : "4",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2020-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_id" : "1",
"_score" : 1.151647,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Mary F.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "yes"
}
}
]
}
}
问题：
- 为什么返回的第一个文档是
approval = no？我期待带有approval = yes的文档会得到更好的评分。 - 为什么索引 = 5 的文档（它不符合条件
category.name，但符合approved = yes）没有被返回？ approved = yes的可选性未在上述查询中表达。我怎样才能用minimum_should_match: 0创建一种额外的分隔should术语？可以增加分数但不会过滤结果的东西。
我用你的映射和示例数据尝试了你的场景，发现了问题，你在导致问题的 nested 查询上下文中使用了 approved:yes，这就是导致问题的原因，如果您将查询更改为下面的 （基本上在 should 块中使用 approved:yes 但在 nested 查询之外），它会解决您所有的问题。
{
"query": {
"bool": {
"should": [
{
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{
"match": {
"category.level": "A"
}
}
],
"should": [
{
"term": {
"category.name": "John G."
}
},
{
"term": {
"category.name": "Chris T."
}
}
]
}
}
}
},
{
"term": {
"approved": "yes"
}
}
]
}
}
}
和搜索结果
"hits": [
{
"_index": "71967271",
"_id": "4",
"_score": 1.9845366,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2020-04-18 19:09:27.527+0200",
"approved": "yes"
}
},
{
"_index": "71967271",
"_id": "2",
"_score": 1.4455402,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Chris T.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "no"
}
},
{
"_index": "71967271",
"_id": "1",
"_score": 1.2437345,
"_source": {
"category": [
{
"name": "John G.",
"level": "A"
},
{
"name": "Mary F.",
"level": "A"
}
],
"createdBy": "John",
"createdAt": "2022-04-18 19:09:27.527+0200",
"approved": "yes"
}
},
{
"_index": "71967271",
"_id": "5",
"_score": 0.7968255,
"_source": {
"category": [
{
"name": "Unknown A.",
"level": "A"
},
{
"name": "Unknown B.",
"level": "A"
}
],
"createdBy": "Unknown",
"createdAt": "2020-08-18 19:09:27.527+0200",
"approved": "yes"
}
}
]
您需要使用下面的查询，其中有主要的 bool 查询。它有第一个带有嵌套查询的 must 子句，它有 bool 查询 category.level 字段，然后另一个 bool 查询带有 should 子句 category.name字段。
现在主 bool 查询有 approved 的 should 子句，用于提升具有 yes 值的结果（这在 nested 查询之外）。
POST data/_search
{
"query": {
"bool": {
"must": [
{
"nested": {
"path": "category",
"query": {
"bool": {
"must": [
{
"term": {
"category.level": {
"value": "a"
}
}
},
{
"bool": {
"should": [
{
"term": {
"category.name": "John G."
}
},
{
"term": {
"category.name": "Chris T."
}
}
]
}
}
]
}
}
}
}
],
"should": [
{
"term": {
"approved": "yes"
}
}
]
}
}
}
结果：
{
"took" : 2,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 3,
"relation" : "eq"
},
"max_score" : 1.9845366,
"hits" : [
{
"_index" : "data",
"_type" : "_doc",
"_id" : "4",
"_score" : 1.9845366,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2020-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.6906434,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Mary F.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "yes"
}
},
{
"_index" : "data",
"_type" : "_doc",
"_id" : "2",
"_score" : 1.4455402,
"_source" : {
"category" : [
{
"name" : "John G.",
"level" : "A"
},
{
"name" : "Chris T.",
"level" : "A"
}
],
"createdBy" : "John",
"createdAt" : "2022-04-18 19:09:27.527+0200",
"approved" : "no"
}
}
]
}
}
Why the first document returned is an approval = no? I was expecting that docs with approval = yes would be better scored.
因为您在 nested 查询中有 should 子句，并且它与任何文档都不匹配，因为 approved 在 category 之外，因此它不会改变分数。
Why doc with index = 5 (it doesn't attend the criteria category.name, but it does for approved = yes) is not being returned?
它已被您的 must 子句删除，但是如果您还需要 index =5 文档，那么您可以添加两个 should 子句，一个用于嵌套，一个用于 approved，它将解决你的问题。
我的回答也解决了你的问题3。