Elasticsearch学习--索引的批量操作mget、bulk

雨点打透心脏的1/2处 2024-04-03 09:59 84阅读 0赞

### 一、批量查询mget ###

#### 1. 基本用法 ####

查询id是1、2的数据

# 批量查询
    GET _mget
    {
      "docs":[
        {
          "_index":"indexname",
          "_id":1
        },
        {
          "_index":"indexname",
          "_id":2
        } 
      ]
    }

![00d2ca6d2acc4cf28c2844f420f24f5a.png][]

#### 2. 提取index ####

# 也可以把索引提到上面
    GET indexname/_mget
    {
      "docs":[
        {
          "_id":1
        },
        {
          "_id":2
        } 
      ]
    }

#### 3. ids的用法 ####

# 也可以通过ids直接查询id
    GET indexname/_mget
    {
      "ids":[1,2]
    }

#### 4. 指定source ####

# 查询时，指定source
    GET indexname/_mget
    {
      "docs":[
        {
          "_id":1,
          "_source":[
            "name"
            ]
          
        },
        {
          "_id":2,
          "_source":{
            "include":[],
            "exclude":[]
          }
        } 
      ]
    }

### 二、文档的操作类型 ###

create：不存在则创建，存在则报错  
delete：删除文档  
update：全量替换或部分更新  
index：索引（动词）

#### 1. 自动生成id ####

# 自动生成id
    POST indexname/_doc
    {
      "a":"3",
      "b":5
    }

![db535ea314d44209830edb0e426af682.png][]

#### 2. 删除操作是懒删除 ####

并没有真正的删除，只是标记为删除

#### 3. index（可以是创建，也可以是全量替换） ####

如果数据不存在，创建：PUT \{indexname\}/\_create/\{id\}

如果数据存在，全量替换：PUT \{indexname\}/\_doc/\{id\}

PUT indexname/_doc/1?op_type=index
    {
      "a":1,
      "b":2
    }

#### 4. filter\_path=items.\*.error ####

请求后面加这个参数，只输出错误信息

PUT indexname/_doc/10?filter_path=items.*.error
    {
      "a":1,
      "b":2
    }

### 三、批量增删改 ###

#### 1. 新增 create ####

POST  _bulk
    {"create":{"_index":"product","_id":10}}
    {"name":"xxxx"}

![899079cb68d645d49cca4f0fc18ad116.png][]

#### 2. 删除 ####

POST  _bulk
    {"delete":{"_index":"product","_id":10}}

![ea7c489a4fab4aaf8156d52b47f3046d.png][]

#### 3. 修改 ####

部分修改，数据不存在会报错

POST  _bulk
    {"update":{"_index":"product","_id":10}}
    {"doc":{"name":"xxxx22"}}

#### 4. 批量操作 ####

也可以把增删改放在一起，批量操作

POST  _bulk
    {"create":{"_index":"product","_id":11}}
    {"name":"xxxx"}
    {"create":{"_index":"product","_id":12}}
    {"name":"xxxx12"}
    {"update":{"_index":"product","_id":12}}
    {"doc":{"name":"xxxx13"}}
    {"delete":{"_index":"product","_id":12}}
    {"delete":{"_index":"product","_id":11}}
    {"delete":{"_index":"product","_id":10}}

![11bf5c80d21f4f56af9d91294402d085.png][]

如果操作数据过多，结果也会有很多，只想看错误结果：

POST  _bulk?filter_path=items.*.error
    {"create":{"_index":"product","_id":11}}
    {"name":"xxxx"}
    {"create":{"_index":"product","_id":12}}
    {"name":"xxxx12"}
    {"update":{"_index":"product","_id":12}}
    {"doc":{"name":"xxxx13"}}
    {"update":{"_index":"product","_id":13}}
    {"doc":{"name":"xxxx13"}}
    {"delete":{"_index":"product","_id":12}}
    {"delete":{"_index":"product","_id":11}}
    {"delete":{"_index":"product","_id":10}}

id为13的数据操作失败

![54e8270c76ba42e9bc26a632bbc4eb49.png][]

#### 5. bulk操作的优缺点 ####

**优点**：不消耗额外的内存。单条操作时，需要把\{\}中的数据在内存中序列化成json对象，消耗堆内存空间

**缺点**：人进行阅读时，可读性较差

[00d2ca6d2acc4cf28c2844f420f24f5a.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/14ecd997abec44328eef063c1c1c8325.png
[db535ea314d44209830edb0e426af682.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/aed75c1192c14de680bac55f74742663.png
[899079cb68d645d49cca4f0fc18ad116.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/0c6afb4f06204823873540cf06fead0e.png
[ea7c489a4fab4aaf8156d52b47f3046d.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/34def8a45901412ca7eb6d3e7644b80f.png
[11bf5c80d21f4f56af9d91294402d085.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/ef0a692644294b68be1d999804fe7d83.png
[54e8270c76ba42e9bc26a632bbc4eb49.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/03/990904ba3f3043d0a47b3709ff624da8.png