原文地址:http://blog.itpub.net/29806344/viewspace-1399621/
在分析sphix原理之前,我先澄清一下为什么经常出现coreseek这个词?
因为sphinx默认不支持中文索引及检索,而coreseek基于sphinx开发了coreseek全文检索服务器,它提供了为sphinx设计的中文分词包libmmseg包含mmseg中文分词,是目前用的最多的sphinx中文检索。
在没有sphinx之前,mysql数据库要对海量的文章中的词进行全文索引,一般用的语句例如:SELECT *** WHERE *** LIKE '%word%';这样的LIKE查询,并且再结合通配符%,是使用不到mysql本身的索引,需要全表扫描,时间超慢的!
如果用到sphinx,全文索引交给sphinx来做,sphinx返回含有该word的ID号,然后用该ID号直接去数据库准确定位那些数据,整个过程如下图:
sphinx的索引文件存储的不是完整的数据,只是由ID和分词组成的数组,由于索引文件不同直接查看,但我们可以通过search工具来验证:
先建索引:
/usr/local/coreseek/bin/indexer -c
/usr/local/coreseek/etc/sphinx.conf
Coreseek Fulltext 4.1 [ Sphinx 2.0.2-dev
(r2922)]
Copyright (c) 2007-2011,
Beijing Choice Software Technologies Inc
(http://www.coreseek.com)
再通过search 查找单词test:
/usr/local/coreseek/bin/search test -c
/usr/local/coreseek/etc/sphinx.conf
Coreseek Fulltext 4.1 [ Sphinx 2.0.2-dev
(r2922)]
Copyright (c) 2007-2011,
Beijing Choice Software Technologies Inc
(http://www.coreseek.com)
using config file '/usr/local/coreseek/etc/sphinx.conf'...
index 'test1': query 'test ': returned 3
matches of 3 total in 0.050 sec
displaying matches:
1. document=1, weight=2421, group_id=1,
date_added=Thu Jan 8 21:43:32 2015
id=1
group_id=1
group_id2=5
date_added=2015-01-08
21:43:32
title=test
one
content=this
is my test document number one. also checking search within phrases.
2. document=2, weight=2421, group_id=1,
date_added=Thu Jan 8 21:43:32 2015
id=2
group_id=1
group_id2=6
date_added=2015-01-08
21:43:32
title=test
two
content=this
is my test document number two
3. document=4, weight=1442, group_id=2,
date_added=Thu Jan 8 21:43:32 2015
id=4
group_id=2
group_id2=8
date_added=2015-01-08
21:43:32
title=doc
number four
content=this
is to test groups
words:
1. 'test': 3 documents, 5 hits
再通过search 查找单词this:
/usr/local/coreseek/bin/search this -c
/usr/local/coreseek/etc/sphinx.conf
Coreseek Fulltext 4.1 [ Sphinx 2.0.2-dev
(r2922)]
Copyright (c) 2007-2011,
Beijing Choice Software Technologies Inc
(http://www.coreseek.com)
using config file
'/usr/local/coreseek/etc/sphinx.conf'...
index 'test1': query 'this ': returned 4
matches of 4 total in 0.000 sec
displaying matches:
1. document=1, weight=1304, group_id=1,
date_added=Thu Jan 8 21:43:32 2015
id=1
group_id=1
group_id2=5
date_added=2015-01-08
21:43:32
title=test
one
content=this
is my test document number one. also checking search within phrases.
2. document=2, weight=1304, group_id=1,
date_added=Thu Jan 8 21:43:32 2015
id=2
group_id=1
group_id2=6
date_added=2015-01-08
21:43:32
title=test
two
content=this
is my test document number two
3. document=3, weight=1304, group_id=2,
date_added=Thu Jan 8 21:43:32 2015
id=3
group_id=2
group_id2=7
date_added=2015-01-08
21:43:32
title=another
doc
content=this
is another group
4. document=4, weight=1304, group_id=2,
date_added=Thu Jan 8 21:43:32 2015
id=4
group_id=2
group_id2=8
date_added=2015-01-08
21:43:32
title=doc
number four
content=this
is to test groups
words:
1. 'this': 4 documents, 4 hits
由此,我们可以看到,search 关键词
主要返回的是含有表ID和命中率的数组。
注意:不知道大家有没有想到一个致命的问题,创建了sphinx全文索引后,如果在mysql中新增加数据,不重新indexer一下,sphinx索引是搜索不到的!即使是加参数–rotate,数据多的情况下,也要很长时间,这个问题怎么解决呢!明天就来讲主索引和增量索引,以及用cron来处理新数据自动加入增量索引中。
补充: