ElasticSearch(四)：ES nested嵌套文档与父子文档处理

对于复杂的嵌套字段处理使用nested来避免数据扁平化处理，使用数据如下：

PUT /user_index
{"mappings": {"properties": {"user_name": {"type": "keyword"},"age ": {"type": "short"},"address": {"properties": {"province": {"type": "keyword"},"city": {"type": "keyword"},"street": {"type": "keyword"}}}}}
}

插入数据


POST _bulk
{"create":{"_index":"user_index", "_type":"_doc", "_id":1}}
{"id":1,"user_name":"jack","age":"25","address":[{"province":"北京","city": "北京","street": "枫林三路"},{"province":"天津","city": "天津","street": "华夏路"}]}
{"create":{"_index":"user_index", "_type":"_doc", "_id":2}}
{"id":2,"user_name":"rose","age":"23","address":[{"province":"河北","city": "廊坊","street": "燕郊经济开发区"},{"province":"天津","city": "天津","street": "华夏路"}]}
{"create":{"_index":"user_index", "_type":"_doc", "_id":3}}
{"id":3,"user_name":"libai","age":"21","address":[{"province":"上海","city": "上海","street": "淞虹路"},{"province":"河南","city": "郑州","street": "莲花路"}]}```

如下根据省份城市查询

GET /user_index/_search
{"query": {"bool": {"must": [{"match": {"address.province": "北京"}},{"match": {"address.city": "天津"}}]}}
}

查询结果：

{"took" : 0,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 1,"relation" : "eq"},"max_score" : 1.8239043,"hits" : [{"_index" : "user_index","_type" : "_doc","_id" : "1","_score" : 1.8239043,"_source" : {"id" : 1,"user_name" : "jack","age" : "25","address" : [{"province" : "北京","city" : "北京","street" : "枫林三路"},{"province" : "天津","city" : "天津","street" : "华夏路"}]}}]}
}

nested 查询

但是得到的结果并不准确，这个时候就需要使用nested object来定义数据建模。

DELETE /user_index
PUT /user_index
{"mappings": {"properties": {"user_name": {"type": "keyword"},"age ": {"type": "short"},"address": {"type": "nested", "properties": {"province": {"type": "keyword"},"city": {"type": "keyword"},"street": {"type": "keyword"}}}}}
}

这个时候使用nested查询

GET /user_index/_search
{"query": {"bool": {"must": [{"nested": {"path": "address","query": {"bool": {"must": [{"match": {"address.province": "北京"}},{"match": {"address.city": "天津"}}]}}}}]}}
}

查询结果如下：

{"took" : 0,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 0,"relation" : "eq"},"max_score" : null,"hits" : [ ]}
}

虽然语法变的复杂了，但是在数据的读写操作上都不会有错误发生，是推荐的设计方式。

其原因是：
普通的数组数据在ES中会被扁平化处理，处理方式如下：（如果字段需要分词，会将分词数据保存在对应的字段位置，当然应该是一个倒排索引，这里只是一个直观的案例）

{"user_name" : "jack","address.province" : [ "北京", "天津" ],"address.city" : [ "北京", "天津" ]"address.street" : [ "枫林三路", "华夏路" ]
}

那么nested object数据类型ES在保存的时候不会有扁平化处理，保存方式如下：所以在搜索的时候一定会有需要的搜索结果。

{"user_name" : "jack"
}
{"address.province" : "北京","address.city" : "北京"，"address.street" : "枫林三路"
}
{"address.province" : "天津","address.city" : "天津","address.street" : "华夏路",
}

nested Object建模的优缺点：
缺点：
1：数据冗余将多个数据都放在一起了，维护成本就比较高
2：每次更新，需要重新索引整个对象（包括跟对象和嵌套对象）
优点：
1：文档存储在一起，读取效率性能高
适用于查询频繁更新较少的场景

父子文档

ES 提供了类似关系型数据库中 Join 的实现。使用 Join 数据类型实现，可以通过 Parent / Child 的关系，从而分离两个对象

父文档和子文档是两个独立的文档，更新父文档无需重新索引整个子文档。子文档被新增，更改和删除也不会影响到父文档和其他子文档。
要点：父子关系元数据映射，用于确保查询时候的高性能，但是有一个限制，就是父子数据必须存在于一个shard中

父子关系

DELETE user_addressPUT user_address
{"mappings": {"properties": {"user_address_relation": {"type": "join","relations": {"user": "address"}},"user_name": {"type": "keyword"},"age ": {"type": "short"}}}
}

插入user父文档

PUT user_address/_doc/user1
{"user_name": "jack","age": "25","user_address_relation": {"name": "user"}
}PUT user_address/_doc/user2
{"user_name": "rose","age": "23","user_address_relation": {"name": "user"}
}

索引子文档

父文档和子文档必须存在相同的分片上，确保查询 join 的性能
当指定文档时候，必须指定它的父文档 ID ，使用 route 参数来保证，分配到相同的分片

PUT user_address/_doc/address1?routing=user1
{"province": "北京","city": "北京","street": "枫林三路","user_address_relation": {"name": "address","parent": "user1"}
}PUT user_address/_doc/address2?routing=user1
{"province": "天津","city": "天津","street": "华夏路","user_address_relation": {"name": "address","parent": "user1"}
}PUT user_address/_doc/address3?routing=user2
{"province": "河北","city": "廊坊","street": "燕郊经济开发区","user_address_relation": {"name": "address","parent": "user2"}
}PUT user_address/_doc/address4?routing=user2
{"province": "天津","city": "天津","street": "华夏路","user_address_relation": {"name": "address","parent": "user2"}
}

Parent / Child 所支持的查询

查询所有文档

POST user_address/_search
{}

{"took" : 372,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 6,"relation" : "eq"},"max_score" : 1.0,"hits" : [{"_index" : "user_address","_type" : "_doc","_id" : "user1","_score" : 1.0,"_source" : {"user_name" : "jack","age" : "25","user_address_relation" : {"name" : "user"}}},{"_index" : "user_address","_type" : "_doc","_id" : "user2","_score" : 1.0,"_source" : {"user_name" : "rose","age" : "23","user_address_relation" : {"name" : "user"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address1","_score" : 1.0,"_routing" : "user1","_source" : {"province" : "北京","city" : "北京","street" : "枫林三路","user_address_relation" : {"name" : "address","parent" : "user1"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address2","_score" : 1.0,"_routing" : "user1","_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user1"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address3","_score" : 1.0,"_routing" : "user2","_source" : {"province" : "河北","city" : "廊坊","street" : "燕郊经济开发区","user_address_relation" : {"name" : "address","parent" : "user2"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address4","_score" : 1.0,"_routing" : "user2","_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user2"}}}]}
}

根据父文档ID查看

GET user_address/_doc/user1

查询结果：

{"_index" : "user_address","_type" : "_doc","_id" : "user1","_version" : 1,"_seq_no" : 0,"_primary_term" : 1,"found" : true,"_source" : {"user_name" : "jack","age" : "25","user_address_relation" : {"name" : "user"}}
}

Parent Id 查询

返回所有相关子文档
通过对付文档 Id 进行查询，返回所有相关的子文档

POST user_address/_search
{"query": {"parent_id": {"type": "address","id": "user2"}}
}

查询结果：

{"took" : 1,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 2,"relation" : "eq"},"max_score" : 0.6931471,"hits" : [{"_index" : "user_address","_type" : "_doc","_id" : "address3","_score" : 0.6931471,"_routing" : "user2","_source" : {"province" : "河北","city" : "廊坊","street" : "燕郊经济开发区","user_address_relation" : {"name" : "address","parent" : "user2"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address4","_score" : 0.6931471,"_routing" : "user2","_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user2"}}}]}
}

Has Child 查询

返回父文档
通过对子文档进行查询，返回具体相关子文档的父文档
父子文档在相同的分片上，因此 Join 效率高


POST user_address/_search
{"query": {"has_child": {"type": "address","query": {"match": {"province": "北京"}}}}
}

查询结果：

{"took" : 2,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 2,"relation" : "eq"},"max_score" : 1.0,"hits" : [{"_index" : "user_address","_type" : "_doc","_id" : "user1","_score" : 1.0,"_source" : {"user_name" : "jack","age" : "25","user_address_relation" : {"name" : "user"}}},{"_index" : "user_address","_type" : "_doc","_id" : "user2","_score" : 1.0,"_source" : {"user_name" : "rose","age" : "23","user_address_relation" : {"name" : "user"}}}]}
}

Has Parent 查询

返回相关性的子文档
通过对父文档进行查询，返回相关的子文档

POST user_address/_search
{"query": {"has_parent": {"parent_type": "user","query": {"match": {"user_name": "jack"}}}}
}

查询结果

{"took" : 3,"timed_out" : false,"_shards" : {"total" : 1,"successful" : 1,"skipped" : 0,"failed" : 0},"hits" : {"total" : {"value" : 2,"relation" : "eq"},"max_score" : 1.0,"hits" : [{"_index" : "user_address","_type" : "_doc","_id" : "address1","_score" : 1.0,"_routing" : "user1","_source" : {"province" : "北京","city" : "北京","street" : "枫林三路","user_address_relation" : {"name" : "address","parent" : "user1"}}},{"_index" : "user_address","_type" : "_doc","_id" : "address2","_score" : 1.0,"_routing" : "user1","_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user1"}}}]}
}

访问子文档

需指定父文档 routing 参数

GET user_address/_doc/address2

查询结果：

{"_index" : "user_address","_type" : "_doc","_id" : "address2","_version" : 1,"_seq_no" : 7,"_primary_term" : 1,"_routing" : "user1","found" : true,"_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user1"}}
}

#通过ID和routing ，访问子文档
GET user_address/_doc/address2?routing=user2

查询结果：

{"_index" : "user_address","_type" : "_doc","_id" : "address2","_version" : 1,"_seq_no" : 7,"_primary_term" : 1,"_routing" : "user1","found" : true,"_source" : {"province" : "天津","city" : "天津","street" : "华夏路","user_address_relation" : {"name" : "address","parent" : "user1"}}
}

更新子文档

更新子文档不会影响到父文档

PUT user_address/_doc/address2?routing=user2
{"province": "上海","city": "上海","street": "南京西路","user_address_relation": {"name": "address","parent": "user2"}
}
再次查询子文档
GET user_address/_doc/address2?routing=user2

查询结果如下：

{"_index" : "user_address","_type" : "_doc","_id" : "address2","_version" : 2,"_seq_no" : 10,"_primary_term" : 1,"_routing" : "user2","found" : true,"_source" : {"province" : "上海","city" : "上海","street" : "南京西路","user_address_relation" : {"name" : "address","parent" : "user2"}}
}

优点：
1:父子文档独立更新互不影响
2:父子关系数据存在一个shard中，而且还有映射其关联关系的元数据，那么搜索父子关系数据的时候，不用跨分片，一个分片本地自己就搞定了，性能当然高
缺点：
1：需要额外的内存去维护关系。读取性能相对差
适合更新频繁的操作

创作不易，如果对你有帮助，帮忙点个赞呗。