MySql索引详解

1、索引

索引是表的目录，在查找内容之前可以先在目录中查找索引位置，以此快速定位查询数据。对于索引，会保存在额外的文件中。

2、索引，是数据库中专门用于帮助用户快速查询数据的一种数据结构。类似于字典中的目录，查找字典内容时可以根据目录查找到数据的存放位置，然后直接获取即可。

索引由数据库中一列或多列组合而成，其作用是提高对表中数据的查询速度
索引的优点是可以提高检索数据的速度
索引的缺点是创建和维护索引需要耗费时间
索引可以提高查询速度，会减慢写入速度

索引分类

1.普通索引
2.唯一索引
3.全文索引
4.单列索引
5.多列索引
6.空间索引
7.主键索引
8.组合索引

• 普通索引：仅加速查询
• 唯一索引：加速查询 + 列值唯一（可以有null）
• 主键索引：加速查询 + 列值唯一 +　表中只有一个（不可以有null）
• 组合索引：多列值组成一个索引，
  专门用于组合搜索，其效率大于索引合并
• 全文索引：对文本的内容进行分词，进行搜索

索引合并，使用多个单列索引组合搜索
覆盖索引，select的数据列只用从索引中就能够取得，不必读取数据行，换句话说查询列要被所建的索引覆盖

如何创建索引？记住一个单词—explain

创建表的时候创建索引

1 2 3 4 5 6

<span class= "hljs-keyword" > CREATE  <span class= "hljs-keyword" > TABLE  tbl_name( 字段名称 字段类型 [完整性约束条件]， ，，，, [<span class= "hljs-keyword" > UNIQUE |FULLTEXT|SPATIAL] <span class= "hljs-keyword" > INDEX |<span class= "hljs-keyword" > KEY [索引名称](字段名称[(长度)]) [<span class= "hljs-keyword" > ASC |<span class= "hljs-keyword" > DESC ] );</span></span></span></span></span></span></span>

在已经存在的表上创建索引：

1 2	1. CREATE  [ UNIQUE |FULLTEXT|SPATIAL] INDEX  索引名称 ON  表名{字段名称[(长度)] [ ASC | DESC ]} 2. ALTER  TABLE  tbl_name ADD  [<span class= "hljs-string" > UNIQUE |FULLTEXT|SPATIAL] INDEX 索引名称(字段名称[<span class= "hljs-string" >(长度)][<span class= "hljs-symbol" > ASC | DESC ]);</span></span></span>

如何删除索引？

1	<span class= "hljs-keyword" > DROP  <span class= "hljs-keyword" > INDEX  索引名称 <span class= "hljs-keyword" > ON  tbl_name</span></span></span>

1.普通索引

普通索引仅有一个功能：加速查询

创建表+索引

1 2 3 4 5 6 7

<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  in1(      nid <span class= "hljs-built_in" > int  <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null  auto_increment primary  <span class= "hljs-keyword" > key ,      <span class= "hljs-keyword" > name  <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >32) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      email <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >64) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      extra <span class= "hljs-built_in" >text,      <span class= "hljs-keyword" > index  ix_name (<span class= "hljs-keyword" > name ) )</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

创建索引

1	<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > index  index_name <span class= "hljs-keyword" > on  table_name(column_name)</span></span></span>

删除索引

1	<span class= "hljs-keyword" > drop  index_name <span class= "hljs-keyword" > on  table_name;</span></span>

查看索引

1	<span class= "hljs-keyword" >show <span class= "hljs-keyword" > index  <span class= "hljs-keyword" > from  table_name;</span></span></span>

注意：对于创建索引时如果是BLOB 和 TEXT 类型，必须指定length。

1	<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > index  ix_extra <span class= "hljs-keyword" > on  in1(extra(<span class= "hljs-number" >32));</span></span></span></span>

2、唯一索引

唯一索引有两个功能：加速查询 和 唯一约束（可含null）

创建表+唯一索引

1 2 3 4 5 6 7

<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  in1(      nid <span class= "hljs-built_in" > int  <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null  auto_increment primary  <span class= "hljs-keyword" > key ,      <span class= "hljs-keyword" > name  <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >32) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      email <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >64) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      extra <span class= "hljs-built_in" >text,      <span class= "hljs-keyword" > unique  ix_name (<span class= "hljs-keyword" > name ) )</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

创建唯一索引

1	<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > unique  <span class= "hljs-keyword" > index  索引名 <span class= "hljs-keyword" > on  表名(列名)</span></span></span></span>

删除唯一索引

1	<span class= "hljs-keyword" > drop  <span class= "hljs-keyword" > unique  <span class= "hljs-keyword" > index  索引名 <span class= "hljs-keyword" > on  表名</span></span></span></span>

3、主键索引

主键有两个功能：加速查询 和 唯一约束（不可含null）

创建表+创建主键

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  in1(      nid <span class= "hljs-built_in" > int  <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null  auto_increment primary  <span class= "hljs-keyword" > key ,      <span class= "hljs-keyword" > name  <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >32) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      email <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >64) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      extra <span class= "hljs-built_in" >text,      <span class= "hljs-keyword" > index  ix_name (<span class= "hljs-keyword" > name ) )   <span class= "hljs-keyword" > OR   <span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  in1(      nid <span class= "hljs-built_in" > int  <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null  auto_increment,      <span class= "hljs-keyword" > name  <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >32) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      email <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >64) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      extra <span class= "hljs-built_in" >text,      primary  <span class= "hljs-keyword" > key (ni1),      <span class= "hljs-keyword" > index  ix_name (<span class= "hljs-keyword" > name ) )</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

创建主键

1	<span class= "hljs-keyword" > alter  <span class= "hljs-keyword" > table  表名 <span class= "hljs-keyword" > add  primary  <span class= "hljs-keyword" > key (列名);</span></span></span></span>

删除主键

1 2

<span class= "hljs-keyword" > alter  <span class= "hljs-keyword" > table  表名 <span class= "hljs-keyword" > drop  primary  <span class= "hljs-keyword" > key ; <span class= "hljs-keyword" > alter  <span class= "hljs-keyword" > table  表名  <span class= "hljs-keyword" > modify   列名 <span class= "hljs-built_in" > int , <span class= "hljs-keyword" > drop  primary  <span class= "hljs-keyword" > key ;</span></span></span></span></span></span></span></span></span></span>

4、组合索引

组合索引是将n个列组合成一个索引

其应用场景为：频繁的同时使用n列来进行查询，如：where n1 = ‘alex’ and n2 = 666。

创建表

1 2 3 4 5 6

<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  in3(      nid <span class= "hljs-built_in" > int  <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null  auto_increment primary  <span class= "hljs-keyword" > key ,      <span class= "hljs-keyword" > name  <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >32) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      email <span class= "hljs-built_in" > varchar (<span class= "hljs-number" >64) <span class= "hljs-keyword" > not  <span class= "hljs-literal" > null ,      extra <span class= "hljs-built_in" >text )</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

创建组合索引

1	<span class= "hljs-keyword" > create  <span class= "hljs-keyword" > index  ix_name_email <span class= "hljs-keyword" > on  in3(<span class= "hljs-keyword" > name ,email);</span></span></span></span>

如上创建组合索引之后，查询：

• name and email  – 使用索引
• name                 — 使用索引
• email                 — 不使用索引

注意：对于同时搜索n个条件时，组合索引的性能好于多个单一索引合并。

相关命令

1 2 3 4 5 6 7 8 9 10 11 12 13

- 查看表结构      desc  表名   - 查看生成表的SQL      <span class= "hljs-keyword" >show <span class= "hljs-keyword" > create  <span class= "hljs-keyword" > table  表名   - 查看索引      <span class= "hljs-keyword" >show <span class= "hljs-keyword" > index  <span class= "hljs-keyword" > from   表名   - 查看执行时间      <span class= "hljs-keyword" > set  profiling = <span class= "hljs-number" >1;      SQL...      <span class= "hljs-keyword" >show <span class= "hljs-keyword" >profiles;</span></span></span></span></span></span></span></span></span></span>

使用索引和不使用索引

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

由于索引是专门用于加速搜索而生，所以加上索引之后，查询效率会快到飞起来。   <span class= "hljs-comment" ># 有索引 mysql> select  * from  tb1 where  name  = <span class= "hljs-string" > 'wupeiqi-888' ; + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-params" >| nid | name         <span class= "hljs-params" >| email               | radom                            <span class= "hljs-params" >| ctime               | + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-params" >| 889 | wupeiqi-<span class= "hljs-number" >888 <span class= "hljs-params" >| wupeiqi888@live.com | <span class= "hljs-number" >5312269e76a16a90b8a8301d5314204b <span class= "hljs-params" >| 2016-08-03 09:33:35 | + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-number" >1 row <span class= "hljs-keyword" > in  set  (<span class= "hljs-number" >0.<span class= "hljs-number" >00 sec)   <span class= "hljs-comment" ># 无索引 mysql> select  * from  tb1 where  email = <span class= "hljs-string" > 'wupeiqi888@live.com' ; + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-params" >| nid | name         <span class= "hljs-params" >| email               | radom                            <span class= "hljs-params" >| ctime               | + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-params" >| 889 | wupeiqi-<span class= "hljs-number" >888 <span class= "hljs-params" >| wupeiqi888@live.com | <span class= "hljs-number" >5312269e76a16a90b8a8301d5314204b <span class= "hljs-params" >| 2016-08-03 09:33:35 | + -----+-------------+---------------------+----------------------------------+---------------------+ <span class= "hljs-number" >1 row <span class= "hljs-keyword" > in  set  (<span class= "hljs-number" >1.23 sec)</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

正确使用索引

数据库表中添加索引后确实会让查询速度起飞，但前提必须是正确的使用索引来查询，如果以错误的方式使用，则即使建立索引也会不奏效。

即使建立索引，索引也不会生效：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

- like  '%xx'      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  <span class= "hljs-keyword" > name  <span class= "hljs-keyword" > like  <span class= "hljs-string" > '%cn' ; - 使用函数      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  <span class= "hljs-keyword" >reverse(<span class= "hljs-keyword" > name ) = <span class= "hljs-string" > 'wupeiqi' ; - or      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid = <span class= "hljs-number" >1 <span class= "hljs-keyword" > or  email = <span class= "hljs-string" > 'seven@live.com' ;      特别的：当 or 条件中有未建立索引的列才失效，以下会走索引              <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid = <span class= "hljs-number" >1 <span class= "hljs-keyword" > or  <span class= "hljs-keyword" > name  = <span class= "hljs-string" > 'seven' ;              <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid = <span class= "hljs-number" >1 <span class= "hljs-keyword" > or  email = <span class= "hljs-string" > 'seven@live.com'  <span class= "hljs-keyword" > and  <span class= "hljs-keyword" > name  = <span class= "hljs-string" > 'alex' - 类型不一致      如果列是字符串类型，传入条件是必须用引号引起来，不然...      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  <span class= "hljs-keyword" > name  = <span class= "hljs-number" >999; - !=      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  <span class= "hljs-keyword" > name  != <span class= "hljs-string" > 'alex'      特别的：如果是主键，则还是会走索引          <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where nid != <span class= "hljs-number" >123 - >      <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  <span class= "hljs-keyword" > name  > <span class= "hljs-string" > 'alex'      特别的：如果是主键或索引是整数类型，则还是会走索引          <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where nid > <span class= "hljs-number" >123          <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where <span class= "hljs-keyword" >num > <span class= "hljs-number" >123 - <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by      <span class= "hljs-keyword" > select  email <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > order <span class= "hljs-keyword" > by  <span class= "hljs-keyword" > name  <span class= "hljs-keyword" > desc ;      当根据索引排序时候，选择的映射如果不是索引，则不走索引      特别的：如果对主键排序，则还是走索引：          <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > order <span class= "hljs-keyword" > by  nid <span class= "hljs-keyword" > desc ;   - 组合索引最左前缀      如果组合索引为：( name ,email)      name  and  email       <span class= "hljs-comment" > -- 使用索引      name                  <span class= "hljs-comment" > -- 使用索引      email                <span class= "hljs-comment" > -- 不使用索引</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

其他注意事项

1 2 3 4 5 6 7 8 9

<span class= "hljs-deletion" >- 避免使用 select  * <span class= "hljs-deletion" >- count (1)或 count (列) 代替 count (*) <span class= "hljs-deletion" >- 创建表时尽量时 char  代替 varchar <span class= "hljs-deletion" >- 表的字段顺序固定长度的字段优先 <span class= "hljs-deletion" >- 组合索引代替多个单列索引（经常使用多个条件查询时） <span class= "hljs-deletion" >- 尽量使用短索引 <span class= "hljs-deletion" >- 使用连接（ JOIN ）来代替子查询(Sub-Queries) <span class= "hljs-deletion" >- 连表时注意条件类型需一致 <span class= "hljs-deletion" >- 索引散列值（重复少）不适合建索引，例：性别不适合</span></span></span></span></span></span></span></span></span>

limit分页

无论是否有索引，limit分页是一个值得关注的问题

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

每页显示10条： 当前 118 120， 125   倒序：              大      小              980    970  7 6  6 5  54  43  32   21 19 98     下一页：        <span class= "hljs-keyword" > select          *      <span class= "hljs-keyword" > from          tb1      <span class= "hljs-keyword" > where          nid < (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid < 当前页最小值 <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  nid <span class= "hljs-keyword" > desc  <span class= "hljs-keyword" >limit 每页数据 *【页码-当前页】) A <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  A.nid <span class= "hljs-keyword" > asc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >1)       <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by          nid <span class= "hljs-keyword" > desc      <span class= "hljs-keyword" >limit <span class= "hljs-number" >10;        <span class= "hljs-keyword" > select          *      <span class= "hljs-keyword" > from          tb1      <span class= "hljs-keyword" > where          nid < (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid < <span class= "hljs-number" >970  <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  nid <span class= "hljs-keyword" > desc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >40) A <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  A.nid <span class= "hljs-keyword" > asc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >1)       <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by          nid <span class= "hljs-keyword" > desc      <span class= "hljs-keyword" >limit <span class= "hljs-number" >10;   上一页：        <span class= "hljs-keyword" > select          *      <span class= "hljs-keyword" > from          tb1      <span class= "hljs-keyword" > where          nid < (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid > 当前页最大值 <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  nid <span class= "hljs-keyword" > asc  <span class= "hljs-keyword" >limit 每页数据 *【当前页-页码】) A <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  A.nid <span class= "hljs-keyword" > asc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >1)       <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by          nid <span class= "hljs-keyword" > desc      <span class= "hljs-keyword" >limit <span class= "hljs-number" >10;        <span class= "hljs-keyword" > select          *      <span class= "hljs-keyword" > from          tb1      <span class= "hljs-keyword" > where          nid < (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid > <span class= "hljs-number" >980 <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  nid <span class= "hljs-keyword" > asc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >20) A <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by  A.nid <span class= "hljs-keyword" > desc  <span class= "hljs-keyword" >limit <span class= "hljs-number" >1)       <span class= "hljs-keyword" > order  <span class= "hljs-keyword" > by          nid <span class= "hljs-keyword" > desc      <span class= "hljs-keyword" >limit <span class= "hljs-number" >10;</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

执行计划

explain + 查询SQL – 用于显示SQL执行信息参数，根据参考信息可以进行SQL优化

1 2 3 4 5 6 7

mysql> explain select  * from  tb2; + ----+-------------+-------+------+---------------+------+---------+------+------+-------+ <span class= "hljs-params" >| id | select_type <span class= "hljs-params" >| table  | type <span class= "hljs-params" >| possible_keys | key   <span class= "hljs-params" >| key_len | ref  <span class= "hljs-params" >| rows  | Extra <span class= "hljs-params" >| + ----+-------------+-------+------+---------------+------+---------+------+------+-------+ |  <span class= "hljs-number" >1 <span class= "hljs-params" >| SIMPLE      | tb2   <span class= "hljs-params" >| ALL   | NULL           <span class= "hljs-params" >| NULL  | NULL     <span class= "hljs-params" >| NULL  |    <span class= "hljs-number" >2 <span class= "hljs-params" >| NULL   | + ----+-------------+-------+------+---------------+------+---------+------+------+-------+ <span class= "hljs-number" >1 row <span class= "hljs-keyword" > in  set  (<span class= "hljs-number" >0.<span class= "hljs-number" >00 sec)</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

执行计划

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

id          查询顺序标识              如：mysql> explain <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid, name  <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid < <span class= "hljs-number" >10) <span class= "hljs-keyword" > as  B;              + ----+-------------+------------+-------+---------------+---------+---------+------+------+-------------+              | id | select_type | table       | type  | possible_keys | key      | key_len | <span class= "hljs-keyword" >ref  | rows  | Extra       |              + ----+-------------+------------+-------+---------------+---------+---------+------+------+-------------+              |  <span class= "hljs-number" >1 | PRIMARY      | <derived2> | ALL    | NULL           | NULL     | NULL     | NULL  |    <span class= "hljs-number" >9 | NULL         |              |  <span class= "hljs-number" >2 | DERIVED     | tb1        | range | PRIMARY        | PRIMARY  | <span class= "hljs-number" >8       | NULL  |    <span class= "hljs-number" >9 | Using <span class= "hljs-keyword" > where  |              + ----+-------------+------------+-------+---------------+---------+---------+------+------+-------------+          特别的：如果使用 union 连接气值可能为<span class= "hljs-literal" > null        select_type          查询类型              SIMPLE          简单查询              PRIMARY          最外层查询              SUBQUERY        映射为子查询              DERIVED         子查询              UNION            联合              UNION  RESULT    使用联合的结果              ...      table          正在访问的表名        type          查询时的访问方式，性能： all  < index  < range < index_merge < ref_or_null < <span class= "hljs-keyword" >ref < eq_ref < system/<span class= "hljs-keyword" >const              ALL              全表扫描，对于数据表从头到尾找一遍                              <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1;                              特别的：如果有limit限制，则找到之后就不在继续向下扫描                                     <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  email = <span class= "hljs-string" > 'seven@live.com'                                     <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  email = <span class= "hljs-string" > 'seven@live.com'  limit <span class= "hljs-number" >1;                                     虽然上述两个语句都会进行全表扫描，第二句使用了limit，则找到一个后就不再继续扫描。                INDEX            全索引扫描，对索引从头到尾找一遍                              <span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1;                RANGE          对索引列进行范围查找                              <span class= "hljs-keyword" > select  *  <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  name  < <span class= "hljs-string" > 'alex' ;                              PS:                                  between  and                                  <span class= "hljs-keyword" > in                                  >   >=  <   <=  操作                                  注意：!= 和 > 符号                INDEX_MERGE     合并索引，使用多个单列索引搜索                              <span class= "hljs-keyword" > select  *  <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  name  = <span class= "hljs-string" > 'alex'  <span class= "hljs-function" > or  nid <span class= "hljs-title" > in  (<span class= "hljs-params" ><span class= "hljs-number" >11,<span class= "hljs-number" >22,<span class= "hljs-number" >33);                REF             根据索引查找一个或多个值                              <span class= "hljs-keyword" > select  *  <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  name  = <span class= "hljs-string" > 'seven' ;                EQ_REF          连接时使用 primary  key  或 unique 类型                              <span class= "hljs-keyword" > select  tb2.nid,tb1. name  <span class= "hljs-keyword" > from  tb2 left  <span class= "hljs-keyword" > join  tb1 on  tb2.nid = tb1.nid;                CONST           常量                              表最多有一个匹配行,因为仅有一行,在这行的列值可被优化器剩余部分认为是常数,<span class= "hljs-keyword" >const表很快,因为它们只读取一次。                              <span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid = <span class= "hljs-number" >2 ;                SYSTEM          系统                              表仅有一行(=系统表)。这是<span class= "hljs-keyword" >const联接类型的一个特例。                              <span class= "hljs-keyword" > select  * <span class= "hljs-keyword" > from  (<span class= "hljs-keyword" > select  nid <span class= "hljs-keyword" > from  tb1 <span class= "hljs-keyword" > where  nid = <span class= "hljs-number" >1) <span class= "hljs-keyword" > as  A;      possible_keys          可能使用的索引        key          真实使用的        key_len          MySQL中使用索引字节长度        rows          mysql估计为了找到所需的行而要读取的行数 ------ 只是预估值        extra          该列包含MySQL解决查询的详细信息          “Using index ”              此值表示mysql将使用覆盖索引，以避免访问表。不要把覆盖索引和 index 访问类型弄混了。          “Using <span class= "hljs-keyword" > where ”              这意味着mysql服务器将在存储引擎检索行后再进行过滤，许多<span class= "hljs-keyword" > where 条件里涉及索引中的列，当（并且如果）它读取索引时，就能被存储引擎检验，因此不是所有带<span class= "hljs-keyword" > where 子句的查询都会显示“Using <span class= "hljs-keyword" > where ”。有时“Using <span class= "hljs-keyword" > where ”的出现就是一个暗示：查询可受益于不同的索引。          “Using temporary ”              这意味着mysql在对查询结果排序时会使用一个临时表。          “Using filesort”              这意味着mysql会对结果使用一个外部索引排序，而不是按索引次序从表里读取行。mysql有两种文件排序算法，这两种排序方式都可以在内存或者磁盘上完成，explain不会告诉你mysql将使用哪一种文件排序，也不会告诉你排序会在内存里还是磁盘上完成。          “<span class= "hljs-function" >Range <span class= "hljs-keyword" >checked <span class= "hljs-keyword" > for  each <span class= "hljs-title" >record(<span class= "hljs-params" > index  map: N)”              这个意味着没有好用的索引，新的索引将在联接的每一行上重新估算，N是显示在possible_keys列中索引的位图，并且是冗余的。</span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>

慢日志查询

a、配置MySQL自动记录慢日志

1 2 3 4

<span class= "hljs-attr" >slow_query_log = <span class= "hljs-literal" > OFF                             是否开启慢日志记录 <span class= "hljs-attr" >long_query_time = <span class= "hljs-number" >2                              时间限制，超过此时间，则记录 <span class= "hljs-attr" >slow_query_log_file = /usr/slow.log        日志文件 <span class= "hljs-attr" >log_queries_not_using_indexes = <span class= "hljs-literal" > OFF      为使用索引的搜索是否记录</span></span></span></span></span></span></span>

注：查看当前配置信息：
　　     show variables like ‘%query%’
     修改当前配置：
　　　　set global 变量名 = 值

b、查看MySQL慢日志

mysqldumpslow -s at -a  /usr/local/var/mysql/MacBook-Pro-3-slow.log

参数详解

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

"" " <span class=" hljs-comment ">--verbose    版本 <span class=" hljs-comment ">--debug      调试 <span class=" hljs-comment ">--help       帮助   -v           版本 -d           调试模式 -s ORDER     排序方式               what to sort by (al, at, ar, c, l, r, t), 'at' is default                al: average <span class=" hljs-keyword ">lock <span class=" hljs-keyword ">time                ar: average <span class=" hljs-keyword ">rows sent                <span class=" hljs-keyword ">at: average <span class=" hljs-keyword ">query <span class=" hljs-keyword ">time                 c: <span class=" hljs-keyword ">count                 l: <span class=" hljs-keyword ">lock <span class=" hljs-keyword ">time                 r: <span class=" hljs-keyword ">rows sent                 t: <span class=" hljs-keyword ">query <span class=" hljs-keyword ">time -r           反转顺序，默认文件倒序拍。<span class=" hljs-keyword ">reverse the <span class=" hljs-keyword ">sort <span class=" hljs-keyword ">order (largest <span class=" hljs-keyword ">last instead <span class=" hljs-keyword ">of <span class=" hljs-keyword ">first) -t <span class=" hljs-keyword ">NUM       显示前N条just <span class=" hljs-keyword ">show the top n queries -a           不要将<span class=" hljs-keyword ">SQL中数字转换成N，字符串转换成S。don<span class=" hljs-string ">'t abstract all numbers to N and strings to 'S<span class=" hljs-string ">' -n NUM       abstract numbers with at least n digits within names -g PATTERN   正则匹配；grep: only consider stmts that include this string -h HOSTNAME  mysql机器名或者IP；hostname of db server for *-slow.log filename (can be wildcard),               default is '*<span class=" hljs-string ">', i.e. match all -i NAME      name of server instance (if using mysql.server startup script) -l           总时间中不减去锁定时间；don't subtract <span class=" hljs-keyword ">lock <span class=" hljs-keyword ">time <span class=" hljs-keyword ">from total <span class=" hljs-keyword ">time <span class=" hljs-string ">" "" </span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span></span>