实战Elasticsearch6的join类型
《Elasticsearch实战》(英文名Elasticsearch IN ACTION)是经典es教程，对应demo源码地址为：https://github.com/dakrone/elasticsearch-in-action ，最新分支6.x，在使用源码时，发现索引_doc的静态映射脚本增加了一个类型为join的字段，如下所示，：
"mappings" : {
"_doc" : {
"_source" : {
"enabled" : true
},
"properties" : {
"relationship_type": {
"type": "join",
"relations" : {
"group": "event"
}
},
...
这是es6新增的类型，一起来通过实战学习这个join；
环境信息
- 操作系统：Ubuntu 18.04.2 LTS
- elasticsearch：6.7.1
- kibana：6.7.1
《Elasticsearch实战》demo源码下载地址
本文用到的源码一共两个文件，一个是创建静态映射的mapping.json，另一个是创建文档的populate.sh，地址分别如下：
- https://github.com/dakrone/elasticsearch-in-action/blob/6.x/mapping.json
- https://github.com/dakrone/elasticsearch-in-action/blob/6.x/populate.sh
上述文件的用法：下载到同一个目录，执行命令./populate.sh 192.168.1.101:9200，"192.168.1.101:9200"是es6的http地址和端口；
官方说法
官方对join类型的说明如下：
我的理解：
- join类型用于建立索引内文档的父子关系；
- 用父子文档的名字来表示关系；
接下来看看《Elasticsearch实战》的demo中是怎么使用这个字段的；
《Elasticsearch实战》的demo
- demo中部分文档的创建脚本如下所示：
curl -s -XPOST "$ADDRESS/get-together/_doc/1" -H'Content-Type: application/json' -d'{
"relationship_type": "group",
"name": "Denver Clojure",
"organizer": ["Daniel", "Lee"],
"description": "Group of Clojure enthusiasts from Denver who want to hack on code together and learn more about Clojure",
"created_on": "2012-06-15",
"tags": ["clojure", "denver", "functional programming", "jvm", "java"],
"members": ["Lee", "Daniel", "Mike"],
"location_group": "Denver, Colorado, USA"
}'
curl -s -XPOST "$ADDRESS/get-together/_doc/100?routing=1" -H'Content-Type: application/json' -d'{
"relationship_type": {
"name": "event",
"parent": "1"
},
"host": ["Lee", "Troy"],
"title": "Liberator and Immutant",
"description": "We will discuss two different frameworks in Clojure for doing different things. Liberator is a ring-compatible web framework based on Erlang Webmachine. Immutant is an all-in-one enterprise application based on JBoss.",
"attendees": ["Lee", "Troy", "Daniel", "Tom"],
"date": "2013-09-05T18:00",
"location_event": {
"name": "Stoneys Full Steam Tavern",
"geolocation": "39.752337,-105.00083"
},
"reviews": 4
}'
如上所示，id为1的记录，其relationship_type字段的值为"group"，id为2的记录，relationship_type字段的值不是字符串，而是对象，parent为1表示父文档id为1，name为"event"表示父子关系是"group:event"类型；
注意：上述第二个文档的地址中携带了routing参数，以保持父子在同一个分片，这是在使用join类型是要格外注意的地方；
接下来，确保前面提到的populate.sh脚本已经执行，使得_doc索引及其文档数据在es环境中准备好，就可以实战了，实战环境是Kibana的Det Tools：
查找所有父类型为"group"的文档(结果是子文档)：
执行如下脚本：
GET get-together/_search
{
"query": {
"has_parent": {
"parent_type": "group",
"query": {
"match_all": {}
}
}
}
}
可以得到所有父类型为"group"的子文档：
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 2,
"successful" : 2,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 15,
"max_score" : 1.0,
"hits" : [
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "106",
"_score" : 1.0,
"_routing" : "3",
"_source" : {
"relationship_type" : {
"name" : "event",
"parent" : "3"
},
"host" : "Mik",
"title" : "Social management and monitoring tools",
"description" : "Shay Banon will be there to answer questions and we can talk about management tools.",
"attendees" : [
"Shay",
"Mik",
"John",
"Chris"
],
"date" : "2013-03-06T18:00",
"location_event" : {
"name" : "Quid Inc",
"geolocation" : "37.798442,-122.399801"
},
"reviews" : 5
}
},
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "107",
"_score" : 1.0,
"_routing" : "3",
"_source" : {
"relationship_type" : {
"name" : "event",
"parent" : "3"
},
"host" : "Mik",
"title" : "Logging and Elasticsearch",
"description" : "Get a deep dive for what Elasticsearch is and how it can be used for logging with Logstash as well as Kibana!",
"attendees" : [
"Shay",
"Rashid",
"Erik",
"Grant",
"Mik"
],
"date" : "2013-04-08T18:00",
"location_event" : {
"name" : "Salesforce headquarters",
"geolocation" : "37.793592,-122.397033"
},
"reviews" : 3
}
},
...
查找所有子类型为"event"的文档（结果是父文档）
执行如下脚本：
GET get-together/_search
{
"query": {
"has_child": {
"type": "event",
"query": {
"match_all": {}
}
}
}
}
可以得到所有子类型为"event"的文档：
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 2,
"successful" : 2,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 5,
"max_score" : 1.0,
"hits" : [
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "3",
"_score" : 1.0,
"_source" : {
"relationship_type" : "group",
"name" : "Elasticsearch San Francisco",
"organizer" : "Mik",
"description" : "Elasticsearch group for ES users of all knowledge levels",
"created_on" : "2012-08-07",
"tags" : [
"elasticsearch",
"big data",
"lucene",
"open source"
],
"members" : [
"Lee",
"Igor"
],
"location_group" : "San Francisco, California, USA"
}
},
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "1",
"_score" : 1.0,
"_source" : {
"relationship_type" : "group",
"name" : "Denver Clojure",
"organizer" : [
"Daniel",
"Lee"
],
"description" : "Group of Clojure enthusiasts from Denver who want to hack on code together and learn more about Clojure",
"created_on" : "2012-06-15",
"tags" : [
"clojure",
"denver",
"functional programming",
"jvm",
"java"
],
"members" : [
"Lee",
"Daniel",
"Mike"
],
"location_group" : "Denver, Colorado, USA"
}
},
...
查找parent的id等于1的子文档
执行如下脚本：
GET get-together/_search
{
"query": {
"parent_id": {
"type": "event",
"id": "1"
}
}
}
可以得到所有parent的id等于1的子文档：
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 2,
"successful" : 2,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 3,
"max_score" : 1.3291359,
"hits" : [
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "100",
"_score" : 1.3291359,
"_routing" : "1",
"_source" : {
"relationship_type" : {
"name" : "event",
"parent" : "1"
},
"host" : [
"Lee",
"Troy"
],
"title" : "Liberator and Immutant",
"description" : "We will discuss two different frameworks in Clojure for doing different things. Liberator is a ring-compatible web framework based on Erlang Webmachine. Immutant is an all-in-one enterprise application based on JBoss.",
"attendees" : [
"Lee",
"Troy",
"Daniel",
"Tom"
],
"date" : "2013-09-05T18:00",
"location_event" : {
"name" : "Stoneys Full Steam Tavern",
"geolocation" : "39.752337,-105.00083"
},
"reviews" : 4
}
},
...
用script_fields简化返回内容
前面的查询，返回的内容是整个_source，如果不需要全部内容，可以用script_fields来简化；
- 查找所有父文档ID等1的的子文档，并且返回内容只有三个字段：父文档ID、子文档ID、子文档title字段：
GET get-together/_search
{
"query": {
"parent_id": {
"type": "event",
"id": "1"
}
},
"script_fields":{
"group_id":{
"script":{
"source":"doc['relationship_type#group']"
}
},"event_id":{
"script":{
"source":"doc['_id']"
}
},
"title":{
"script":"params['_source']['title']"
}
}
}
得到结果如下：
{
"took" : 1,
"timed_out" : false,
"_shards" : {
"total" : 2,
"successful" : 2,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : 3,
"max_score" : 1.3291359,
"hits" : [
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "100",
"_score" : 1.3291359,
"_routing" : "1",
"fields" : {
"event_id" : [
"100"
],
"title" : [
"Liberator and Immutant"
],
"group_id" : [
"1"
]
}
},
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "101",
"_score" : 1.3291359,
"_routing" : "1",
"fields" : {
"event_id" : [
"101"
],
"title" : [
"Sunday, Surly Sunday"
],
"group_id" : [
"1"
]
}
},
{
"_index" : "get-together",
"_type" : "_doc",
"_id" : "102",
"_score" : 1.3291359,
"_routing" : "1",
"fields" : {
"event_id" : [
"102"
],
"title" : [
"10 Clojure coding techniques you should know, and project openbike"
],
"group_id" : [
"1"
]
}
}
]
}
}
聚合
执行以下查询，会将所有父文档为group的子文档做桶聚合聚合：
GET get-together/_search
{
"query": {
"has_parent": {
"parent_type": "group",
"query": {
"match_all": {}
}
}
},
"aggs":{
"parents":{
"terms":{
"field":"relationship_type#group"
}
}
}
}
得到的结果如下，按照父文档ID得到聚合结果：
"aggregations" : {
"parents" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "1",
"doc_count" : 3
},
{
"key" : "2",
"doc_count" : 3
},
{
"key" : "3",
"doc_count" : 3
},
{
"key" : "4",
"doc_count" : 3
},
{
"key" : "5",
"doc_count" : 3
}
]
}
}
}
以上就是join类型的主要实战内容了，希望能帮助您理解这个新的类型；
欢迎关注我的公众号：程序员欣宸
