前言

本文隶属于专栏《1000个问题搞定大数据技术体系》,该专栏为笔者原创,引用请注明来源,不足和错误之处请在评论区帮忙指出,谢谢!

本专栏目录结构和文献引用请见1000个问题搞定大数据技术体系

正文

1、导入数据

DELETE us
POST /_bulk
{ "create": { "_index": "us", "_type": "tweet", "_id": "1" }}
{ "email" : "john@smith.com", "name" : "John Smith", "username" : "@john" }
{ "create": { "_index": "us", "_type": "tweet", "_id": "2" }}
{ "email" : "mary@jones.com", "name" : "Mary Jones", "username" : "@mary" }
{ "create": { "_index": "us", "_type": "tweet", "_id": "3" }}
{ "date" : "2014-09-13", "name" : "Mary Jones", "tweet" : "Elasticsearch means full text search has never been so easy", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "4" }}
{ "date" : "2014-09-14", "name" : "John Smith", "tweet" : "@mary it is not just text, it does everything", "user_id" : 1 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "5" }}
{ "date" : "2014-09-15", "name" : "Mary Jones", "tweet" : "However did I manage before Elasticsearch?", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "6" }}
{ "date" : "2014-09-16", "name" : "John Smith",  "tweet" : "The Elasticsearch API is really easy to use", "user_id" : 1 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "7" }}
{ "date" : "2014-09-17", "name" : "Mary Jones", "tweet" : "The Query DSL is really powerful and flexible", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "8" }}
{ "date" : "2014-09-18", "name" : "John Smith", "user_id" : 1 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "9" }}
{ "date" : "2014-09-19", "name" : "Mary Jones", "tweet" : "Geo-location aggregations are really cool", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "10" }}
{ "date" : "2014-09-20", "name" : "John Smith", "tweet" : "Elasticsearch surely is one of the hottest new NoSQL products", "user_id" : 1 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "11" }}
{ "date" : "2014-09-21", "name" : "Mary Jones", "tweet" : "Elasticsearch is built for the cloud, easy to scale", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "12" }}
{ "date" : "2014-09-22", "name" : "John Smith", "tweet" : "Elasticsearch and I have left the honeymoon stage, and I still love her.", "user_id" : 1 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "13" }}
{ "date" : "2014-09-23", "name" : "Mary Jones", "tweet" : "So yes, I am an Elasticsearch fanboy", "user_id" : 2 }
{ "create": { "_index": "us", "_type": "tweet", "_id": "14" }}
{ "date" : "2014-09-24", "name" : "John Smith", "tweet" : "How many more cheesy tweets do I have to write?", "user_id" : 1 }

2、size+from浅分页

按照一般的查询流程来说,如果我想查询前10条数据:

  1. 客户端请求发给某个节点
  2. 节点转发给个个分片,查询每个分片上的前10条
  3. 结果返回给节点,整合数据,提取前10条
  4. 返回给请求客户端

from定义了目标数据的偏移值,size定义当前返回的事件数目

GET /us/_search?pretty
{
  "from" : 0 , "size" : 5
}

GET /us/_search?pretty
{
  "from" : 5 , "size" : 5
}

这种浅分页只适合少量数据,因为随from增大,查询的时间就会越大,而且数据量越大,查询的效率指数下降

优点

from+size在数据量不大的情况下,效率比较高

缺点

在数据量非常大的情况下,from+size分页会把全部记录加载到内存中,这样做不但运行速递特别慢,而且容易让es出现内存不足而挂掉

3、scroll深分页

对于上面介绍的浅分页,当Elasticsearch响应请求时,它必须确定docs的顺序,排列响应结果。

如果请求的页数较少(假设每页20个docs), Elasticsearch不会有什么问题,但是如果页数较大时,比如请求第20页,Elasticsearch不得不取出第1页到第20页的所有docs,再去除第1页到第19页的docs,得到第20页的docs。

解决的方式就是使用scroll,scroll就是维护了当前索引段的一份快照信息–缓存(这个快照信息是你执行这个scroll查询时的快照)

可以把 scroll 分为初始化和遍历两步:

  1. 初始化时将所有符合搜索条件的搜索结果缓存起来,可以想象成快照;
  2. 遍历时,从这个快照里取数据;

初始化

GET us/_search?scroll=3m
{ 
"query": {"match_all": {}},
 "size": 3
}

初始化的时候就像是普通的search一样
其中的scroll=3m代表当前查询的数据缓存3分钟
Size:3 代表当前查询3条数据

遍历

在遍历时候,拿到上一次遍历中的scrollid,然后带scroll参数,重复上一次的遍历步骤,知道返回的数据为空,就表示遍历完成

GET /_search/scroll
{
  "scroll" : "1m",
  "scroll_id" : "DnF1ZXJ5VGhlbkZldGNoBQAAAAAAAAPXFk0xN1BmSnlVUldhYThEdWVzZ19xbkEAAAAAAAAAIxZuQWVJU0VSZ1JzcVZtMGVYZ3RDaFlBAAAAAAAAA9oWTVZOdHJ2cXBSOU9wN3c1dk5vcWd4QQAAAAAAAAPYFk0xN1BmSnlVUldhYThEdWVzZ19xbkEAAAAAAAAAIhZuQWVJU0VSZ1JzcVZtMGVYZ3RDaFlB"
}

【注意】:每次都要传参数scroll,刷新搜索结果的缓存时间,另外不需要指定index和type(不要把缓存的时时间设置太长,占用内存)

对比

浅分页

每次查询都会去索引库(本地文件夹)中查询pageNum*page条数据,然后截取掉前面的数据,留下最后的数据。
这样的操作在每个分片上都会执行,最后会将多个分片的数据合并到一起,再次排序,截取需要的。

深分页

可以一次性将所有满足查询条件的数据,都放到内存中。分页的时候,在内存中查询。相对浅分页,就可以避免多次读取磁盘。

上一篇 下一篇