本文共 4666 字，大约阅读时间需要 15 分钟。
1、bulk相当于数据库里的bash操作。
2、引入批量操作bulk，提高工作效率，你想啊，一批一批添加与一条一条添加，谁快?
3、bulk API可以帮助我们同时执行多个请求
4、bulk的格式：
action：index/create/update/delete
metadata：_index,_type,_id
request body：_source (删除操作不需要加request body)
{ action: { metadata }}
{ request body }
5、bulk里为什么不支持get呢？
答：批量操作，里面放get操作，没啥用！所以，官方也不支持。
6、create 和index的区别
如果数据存在，使用create操作失败，会提示文档已经存在，使用index则可以成功执行。
7、bulk一次最大处理多少数据量？
bulk会把将要处理的数据载入内存中，所以数据量是有限制的，最佳的数据量不是一个确定的数值，它取决于你的硬件，你的文档大小以及复杂性，你的索引以及搜索的负载。
一般建议是1000-5000个文档，如果你的文档很大，可以适当减少队列，大小建议是5-15MB，默认不能超过100M，可以在es的配置文件（即$ES_HOME下的config下的elasticsearch.yml）中。
来修改这个值http.max_content_length: 100mb【不建议修改，太大的话bulk也会慢】，
批量操作bulk例子
（1） 比如，我这里，在$ES_HOME里，新建一文件，命名为request。（这里为什么命名为request，去看官网就是）在Linux里，有无后缀没区别。
[hadoop@djt002 elasticsearch-2.4.3]$ pwd
/usr/local/elasticsearch/elasticsearch-2.4.3 [hadoop@djt002 elasticsearch-2.4.3]$ ll total 56 drwxrwxr-x. 2 hadoop hadoop 4096 Feb 20 22:54 bin drwxrwxr-x. 3 hadoop hadoop 4096 Feb 21 01:28 config drwxrwxr-x. 3 hadoop hadoop 4096 Feb 20 22:59 data drwxrwxr-x. 2 hadoop hadoop 4096 Feb 20 22:54 lib -rw-rw-r--. 1 hadoop hadoop 11358 Aug 24 00:46 LICENSE.txt drwxrwxr-x. 2 hadoop hadoop 4096 Feb 21 00:33 logs drwxrwxr-x. 5 hadoop hadoop 4096 Dec 8 00:41 modules -rw-rw-r--. 1 hadoop hadoop 150 Aug 24 00:46 NOTICE.txt drwxrwxr-x. 2 hadoop hadoop 4096 Feb 20 22:59 plugins -rw-rw-r--. 1 hadoop hadoop 8700 Aug 24 00:46 README.textile [hadoop@djt002 elasticsearch-2.4.3]$ vim request [hadoop@djt002 elasticsearch-2.4.3]$ more request {"index":{"_index":"zhouls","_type":"emp","_id":"10"}}{ "name":"jack", "age" :18}{"index":{"_index":"zhouls","_type":"emp","_id":"11"}}{"name":"tom", "age":27}{"update":{"_index":"zhouls","_type":"emp", "_id":"2"}}{"doc":{"age" :22}}{"delete":{"_index":"zhouls","_type":"emp","_id":"1"}} [hadoop@djt002 elasticsearch-2.4.3]$
或者
{ "index" : {"_index":"zhouls","_type":"emp","_id":"21"}}
{ "name" : "test21"}
例子：
{ "index" : { "_index" : "zhouls", "_type" : "type1", "_id" : "1" } }
{ "field1" : "value1" }
{ "index" : { "_index" : "zhouls", "_type" : "type1", "_id" : "2" } }
{ "field1" : "value1" }
{ "delete" : { "_index" : "zhouls", "_type" : "type1", "_id" : "2" } } (删除操作不需要加request body)
{ "create" : { "_index" : "zhouls", "_type" : "type1", "_id" : "3" } }
{ "field1" : "value3" }
{ "update" : {"_index" : "zhouls", "_type" : "type1","_id" : "1" } }
{ "doc" : {"field2" : "value2"} }
（2）使用文件的方式
vi requests
写入批量操作语句。比如，下面
{"index":{"_index":"zhouls","_type":"emp","_id":"10"}}
{ "name":"jack", "age" :18} {"index":{"_index":"zhouls","_type":"emp","_id":"11"}} {"name":"tom", "age":27} {"update":{"_index":"zhouls","_type":"emp", "_id":"2"}} {"doc":{"age" :22}} {"delete":{"_index":"zhouls","_type":"emp","_id":"1"}}
在$ES_HOME目录下，执行下面命令
curl -PUT '192.168.80.200:9200/_bulk' --data-binary @request;
或
curl -XPOST '192.168.80.200:9200/_bulk' --data-binary @request;
[hadoop@djt002 elasticsearch-2.4.3]$ curl -PUT '192.168.80.200:9200/_bulk' --data-binary @request;
{"took":123,"errors":true,"items":[{"index":{"_index":"zhouls","_type":"emp","_id":"10","_version":1,"_shards":{"total":2,"successful":1,"failed":0},"status":201}},{"index":{"_index":"zhouls","_type":"emp","_id":"11","_version":1,"_shards":{"total":2,"successful":1,"failed":0},"status":201}},{"update":{"_index":"zhouls","_type":"emp","_id":"2","status":404,"error":{"type":"document_missing_exception","reason":"[emp][2]: document missing","index":"zhouls","shard":"-1"}}},{"delete":{"_index":"zhouls","_type":"emp","_id":"1","_version":1,"_shards":{"total":2,"successful":1,"failed":0},"status":404,"found":false}}]}[hadoop@djt002 elasticsearch-2.4.3]$
之后，再查看下。
[hadoop@djt002 elasticsearch-2.4.3]$ curl -XGET 'http://192.168.80.200:9200/zhouls/emp/1?pretty'
{ "_index" : "zhouls", "_type" : "emp", "_id" : "1", "found" : false } [hadoop@djt002 elasticsearch-2.4.3]$ curl -XGET 'http://192.168.80.200:9200/zhouls/emp/2?pretty' { "_index" : "zhouls", "_type" : "emp", "_id" : "2", "found" : false } [hadoop@djt002 elasticsearch-2.4.3]$ curl -XGET 'http://192.168.80.200:9200/zhouls/emp/11?pretty' { "_index" : "zhouls", "_type" : "emp", "_id" : "11", "_version" : 4, "found" : true, "_source" : { "name" : "tom", "age" : 27 } }[hadoop@djt002 elasticsearch-2.4.3]$ curl -XGET 'http://192.168.80.200:9200/zhouls/emp/10?pretty'
{ "_index" : "zhouls", "_type" : "emp", "_id" : "10", "_version" : 4, "found" : true, "_source" : { "name" : "jack", "age" : 18 } }
（3） bulk请求可以在URL中声明/_index 或者/_index/_type
这个，自行去测试！
官网
https://www.elastic.co/guide/en/elasticsearch/reference/current/docs-bulk.html
本文转自大数据躺过的坑博客园博客，原文链接：http://www.cnblogs.com/zlslch/p/6422055.html，如需转载请自行联系原作者