MyBatis中MySQL索引失效问题分析及优化策略，避免查询超时现象

mybaitis书写sql需要特别注意where条件中的语句&＃xff0c;否则将会导致索引失效&＃xff0c;使得查询总是超时。如下语句会出现导致索引失效的情况&＃xff1a;

with test1 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_REQUEST a

where a.C_FUNDCODE &＃61; #{cFundcode} and a.D_DATADATE &＃61; #{dDatadate}),

test2 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_CONFIRM b

where b.C_FUNDCODE &＃61; #{cFundcode} and b.D_DATADATE &＃61; #{dDatadate}),

test3 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_DIVIDENDDETAIL c

where c.C_FUNDCODE &＃61; #{cFundcode} and c.D_DATADATE &＃61; #{dDatadate}),

test4 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_SHAREDETAIL f

where f.C_FUNDCODE &＃61; #{cFundcode} and f.D_DATADATE &＃61; #{dDatadate})

select test1.val requestCount,test2.val confirmCount,test3.val dividendCount,test4.val shareDeatilCount from test1,test2,test3,test4 where test1.v&＃61;test2.v and test2.v&＃61;test3.v and test3.v&＃61;test4.v

原因&＃xff1a;直接使用#{dDatadate}导致索引的失效。

sql语句中出现几种情况会导致索引失效&＃xff1a;

1.TO_CHAR(a.D_DATADATE, &＃39;yyyy-mm-dd&＃39;) <&＃61; TO_CHAR(#{dDatadateStart}, &＃39;yyyy-mm-dd&＃39;)&＃xff0c;导致索引失效。

2.trunc(created)>&＃61;TO_DATE(&＃39;2013-12-14&＃39;, &＃39;YYYY-MM-DD&＃39;)&＃xff0c;导致索引失效。

3.c.D_DATADATE &＃61; #{dDatadate}&＃xff0c;导致索引失效。

将上述语句加上TO_DATE函数&＃xff0c;改为如下语句&＃xff0c;不会导致索引失效&＃xff1a;

with test1 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_REQUEST a

where a.C_FUNDCODE &＃61; #{cFundcode} and a.D_DATADATE &＃61; TO_DATE(#{dDatadate},&＃39;yyyy-MM-dd&＃39;)),

test2 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_CONFIRM b

where b.C_FUNDCODE &＃61; #{cFundcode} and b.D_DATADATE &＃61; TO_DATE(#{dDatadate},&＃39;yyyy-MM-dd&＃39;)),

test3 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_DIVIDENDDETAIL c

where c.C_FUNDCODE &＃61; #{cFundcode} and c.D_DATADATE &＃61; TO_DATE(#{dDatadate},&＃39;yyyy-MM-dd&＃39;)),

test4 as (select count(C_FUNDACCO) val,&＃39;a&＃39; v from TINF_SHAREDETAIL f

where f.C_FUNDCODE &＃61; #{cFundcode} and f.D_DATADATE &＃61; TO_DATE(#{dDatadate},&＃39;yyyy-MM-dd&＃39;))

select test1.val requestCount,test2.val confirmCount,test3.val dividendCount,test4.val shareDeatilCount from test1,test2,test3,test4 where test1.v&＃61;test2.v and test2.v&＃61;test3.v and test3.v&＃61;test4.v

百万、千万级别的数据&＃xff0c;加上索引并且使索引生效的sql语句&＃xff0c;查询性能会很快。如果sql使得索引失效&＃xff0c;将会总是超时&＃xff0c;无法加载出来数据。

因此书写sql一定要注意自己写的sql是否使得索引生效&＃xff0c;查看执行计划

FULL全表扫描即未使用索引(或索引失效)&＃xff1b;INDEX使用索引查询&＃xff0c;会使得查询效率非常高。

总结&＃xff1a;百万、千万级别的数据&＃xff0c;在查询sql上使用索引查询&＃xff0c;会很大的提高性能。

原来表中不加索引的查询需要5分钟甚至更长时间导致查询超时&＃xff1b;加上索引后并且sql使用索引生效&＃xff0c;会使得查询几秒就能查询出来数据。

所以利用好索引&＃xff0c;即使面对百万、千万、上亿条数据也不用担心查询超时。