Sphinx是一个基于SQL的全文检索引擎,可以结合MySQL,PostgreSQL做全文搜索,提供比数据库本身更专业的搜索功能特别为mysql也设计了一个存储引擎插件,从此抛弃模糊查询吧。
Sphinx 单一索引最大可包含1亿条记录,在1千万条记录情况下的查询速度为0.x秒(毫秒级)。Sphinx创建100万条记录的索引只要 3、4分钟,创建1000万条记录的索引可以在50分钟内完成,而重建一次只包含最新10万条记录的增量索引只需几十秒。
一、安装
环境:centos6.5
yum install sphinx -y
默认配置路径 /etc/sphinx/ ,在该路径下,有配置文件sphinx.conf,看看我的配置
# 数据源,这里配置的是mysql source src1 { type = mysql sql_host = localhost sql_user = root sql_pass = sql_db = beego_blog sql_port = 3306 # optional, default is 3306 # 创建索引时候,从数据库查询数据的SQL sql_query = \ SELECT id, userid, UNIX_TIMESTAMP(posttime) AS posttime, title, content, tags \ FROM tb_post sql_attr_uint = userid sql_attr_timestamp = posttime sql_query_info = SELECT * FROM tb_post WHERE id=$id } # 索引1 index test1 { # 指定数据源 source = src1 # 索引文件路径 path = /var/lib/sphinx/test1 # 储文档信息的方式 extern docinfo = extern charset_type = sbcs } # 索引2 index testrt { type = rt rt_mem_limit = 32M path = /var/lib/sphinx/testrt charset_type = utf-8 rt_field = title rt_field = content rt_attr_uint = userid } indexer { mem_limit = 32M } searchd { listen = 0.0.0.0:9312 # 索引对外提供服务的地址 listen = 9306:mysql41 log = /var/log/sphinx/searchd.log query_log = /var/log/sphinx/query.log read_timeout = 5 max_children = 30 pid_file = /var/run/sphinx/searchd.pid max_matches = 1000 seamless_rotate = 1 preopen_indexes = 1 unlink_old = 1 workers = threads # for RT to work binlog_path = /var/lib/sphinx }
生成索引,这里我们用上文配置的索引名称test1来从mysql获取数据。因此,我们先在myslq中,创建表和数据
CREATE TABLE `tb_post` ( `id` mediumint(8) unsigned NOT NULL AUTO_INCREMENT, `userid` mediumint(8) unsigned NOT NULL DEFAULT '0' COMMENT '用户id', `title` varchar(100) NOT NULL DEFAULT '' COMMENT '标题', `content` mediumtext NOT NULL COMMENT '内容', `tags` varchar(100) NOT NULL DEFAULT '' COMMENT '标签', `posttime` datetime NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '发布时间', PRIMARY KEY (`id`) ); INSERT INTO `tb_post` VALUES ('1', '1', 'epoll边沿触发漏报消息包问题', '开发一个即时通讯后台,底层的网络收发使用 epoll + main loop实现网络事件', ',技术,', '2016-08-05 11:50:02'); INSERT INTO `tb_post` VALUES ('2', '1', 'epoll 边沿触发和水平触发区别实战讲解', 'epoll,看结果发现只接入了两条,还有3条没接入。说明高并发时,会出现客户端连接不上的问题。', ',技术,', '2016-08-05 22:03:23'); INSERT INTO `tb_post` VALUES ('3', '1', '快速排序算法', '快速排序算法是一个挺经典的算法,值得我们学习', ',技术,', '2016-08-05 23:08:00');
创建索引
[root@centos6 data]# indexer test1 Sphinx 2.0.8-id64-release (r3831) Copyright (c) 2001-2012, Andrew Aksyonoff Copyright (c) 2008-2012, Sphinx Technologies Inc (http://sphinxsearch.com) using config file '/etc/sphinx/sphinx.conf'... indexing index 'test1'... collected 37 docs, 0.8 MB sorted 0.1 Mhits, 100.0% done total 37 docs, 833156 bytes total 0.082 sec, 10061176 bytes/sec, 446.81 docs/sec total 3 reads, 0.000 sec, 57.7 kb/call avg, 0.0 msec/call avg total 9 writes, 0.000 sec, 40.2 kb/call avg, 0.0 msec/call avg
可以看索引了37条文档,我们可以在命令行测试下效果
[root@centos6 libertyblog]# search epoll|more Sphinx 2.0.8-id64-release (r3831) Copyright (c) 2001-2012, Andrew Aksyonoff Copyright (c) 2008-2012, Sphinx Technologies Inc (http://sphinxsearch.com) using config file '/etc/sphinx/sphinx.conf'... index 'test1': query 'epoll ': returned 2 matches of 2 total in 0.000 sec displaying matches: 1. document=59, weight=2831, userid=1, posttime=Fri Aug 5 22:03:23 2016 id=59 userid=1 title=epoll ??????????????? content=开发一个即时通讯后台,底层的网络收发使用 epoll + main loop实现网络事件 ......
结果匹配到了两条数据,篇幅有限,就不全列出来了。数据 1. document=59, weight=2831 表示该索引文档编号59,权重2831。以上是命令行操作,如果我们要对外提供服务,还需要启动searchd服务进程
[root@centos6 data]# service searchd start 正在启动 searchd:Sphinx 2.0.8-id64-release (r3831) Copyright (c) 2001-2012, Andrew Aksyonoff Copyright (c) 2008-2012, Sphinx Technologies Inc (http://sphinxsearch.com) using config file '/etc/sphinx/sphinx.conf'... WARNING: compat_sphinxql_magics=1 is deprecated; please update your application and config listening on 127.0.0.1:9312 listening on all interfaces, port=9306 precaching index 'test1' precaching index 'testrt' precached 2 indexes in 0.002 sec
启动成功,绑定了端口9312,我们查看下状态
[root@centos6 data]# searchd --status Sphinx 2.0.8-id64-release (r3831) Copyright (c) 2001-2012, Andrew Aksyonoff Copyright (c) 2008-2012, Sphinx Technologies Inc (http://sphinxsearch.com) using config file '/etc/sphinx/sphinx.conf'... searchd status -------------- uptime: 252 connections: 1 maxed_out: 0 command_search: 0 command_excerpt: 0 command_update: 0 command_keywords: 0 command_persist: 0 command_status: 1 command_flushattrs: 0 agent_connect: 0 agent_retry: 0 queries: 0 dist_queries: 0 query_wall: 0.000 query_cpu: OFF dist_wall: 0.000 dist_local: 0.000 dist_wait: 0.000 query_reads: OFF query_readkb: OFF query_readtime: OFF avg_query_wall: 0.000 avg_query_cpu: OFF avg_dist_wall: 0.000 avg_dist_local: 0.000 avg_dist_wait: 0.000 avg_query_reads: OFF avg_query_readkb: OFF avg_query_readtime: OFF
现在我们用一个第三方客户端访问该服务(golang语言开发)
package main import ( "github.com/yunge/sphinx" "log" ) func main() { SphinxClient := sphinx.NewClient().SetServer("localhost", 0).SetConnectTimeout(5000) if err := SphinxClient.Error(); err != nil { log.Fatal(err) return } // 查询,第一个参数是我们要查询的关键字,第二个是索引名称test1,第三个是备注 res, err := SphinxClient.Query("epoll", "test1", "search article!") if err != nil { log.Fatal(err) return } var article_ids string for _, match := range res.Matches { article_ids += fmt.Sprintf("%d,", match.DocId) } log.Println(article_ids) SphinxClient.Close() }
打印结果,是 { 1 2 } ,这两个id,没有id为3的,说明索引查找是准确的,因为3里面没有epoll这个单词,而1和2里面都有epoll。至此,我们的测试完成,可以把此功能和自己网站的搜索框对接,以前都是用模糊查询的方式,在数据库中 like ‘%’ 某某,这样效率其实很低,数据多的时候要等半天,现在用第三方索引来实现,速度快好几个量级。
如果有新的数据插入,或者更新数据,是需要做 增量索引 的,很简单
[root@centos6 data]# indexer --rotate test1 Sphinx 2.0.8-id64-release (r3831) Copyright (c) 2001-2012, Andrew Aksyonoff Copyright (c) 2008-2012, Sphinx Technologies Inc (http://sphinxsearch.com) using config file '/etc/sphinx/sphinx.conf'... indexing index 'test1'... collected 37 docs, 0.8 MB sorted 0.1 Mhits, 100.0% done total 37 docs, 833156 bytes total 0.081 sec, 10184036 bytes/sec, 452.26 docs/sec total 3 reads, 0.000 sec, 57.7 kb/call avg, 0.0 msec/call avg total 9 writes, 0.000 sec, 40.2 kb/call avg, 0.1 msec/call avg rotating indices: successfully sent SIGHUP to searchd (pid=12074).
最好把增量索引的操作放到crontab中,定时做增量,以保持索引最新。以下是每天2点做一次增量索引
0 2 * * * indexer --rotate test1
创建于 2015-08-10 杭州,更新于 2016-08-19 杭州
该文章在以下平台同步
- >LIBERALMAN:http://api.liberalman.cn:40000/article/69
- >CSDN:http://blog.csdn.net/socho/article/details/52251177
- >简书:http://www.jianshu.com/p/5dff17e2da7b
- [ ] 引用
via。http://blog.csdn.net/socho/article/details/52251177