批量操作bulk

介绍bulk批量操作API

为了实现多个文档的create、index、update或delete。

官网文档：Bulk API | Elasticsearch Reference [6.2] | Elastic

请求体格式

#请求体格式如下：

{action：{ metadata}}\n

{request body}\n

{action:{metadata}}\n

{request body}\n

...

#\n是换行的转义符号，bulk通过换行识别每一个请求数据，不换行bulk无法区分。json数据不可美化展示。（如图：不能美化展示），因为美化展示的话，就会有多余的换行，bulk无法识别。

为什么是这样的请求体格式：和Elasticsearch的架构设计有关，Elasticsearch的数据是分片的，分片分布在每个节点上，如果请求的数据不换行，会消耗很大的资源解析数据，如果批量操作的数据很大，所需要的内存和cpu资源会很高，不符合Elasticsearch实时快速的特点，Elasticsearch通过换行，在网络缓冲区中，一行一行读取数据，能够更加快速和低消耗的处理每个请求

action

例：{"delete":{"_index":"library","_type":"books","_id":"1"}}

不能美化展示、

#多重模式

#批量操作bulk

POST /library/books/_bulk

{ "index": {"_id":1}}

{"title":"Elasticsearch: The Definitive Guide","price":5}

{"index": {"_id":2}}

{"title":"The Elasticsearch cookbook","price":15}

{"index":{"_id":3}}

{"title":"Elasticsearch Blueprints","price":9}

{"index":{"_id":4}}

{"title":"Thinking in Python","price":22}

{"index":{"_id":5}}

{"title":"Thinking in Java","price":7}

GET /library/

GET /library/books/_mget

{

"ids": ["1","2","3","4","5"]

}

#当然还可以有delete，update等操作

#注意delete下面没有具体的request body

POST /library/books/_bulk

{"delete":{"_index":"library","_type":"books","_id":"1"}}

{"create":{"_index":"music","_type":"classical","_id": "1"}}

{"title": "Ave Verum Corpus" }

{"index":{"_index": "music","_type": "classical"}}

{"title":"Litaniac de Venerabili Altaris Sacromento"}

{"update": {"_index": "library","_type":"books","_id":"2"}}

{"doc":{"price":"18"}}

GET /library/books/1

GET /library/books/_mget

{

"ids": ["1","2","3","4","5"]

}