Preface
 
    There're various data type in MySQL such as number,string,date,time,lob,etc.The data type will convert implicitly if we don't use a proper operand in a SQL statement which may even contain expressions.These implicit conversion of data type sometimes may lead to performance issues because it will make the indexes on the very columns be useless.We should do our best to avoid implicit conversion in our product environment by explicitly specifying the same data type with the relevant column.
    The principle of implicit conversion in MySQL is to make the operation be as compatible as possible and not return result(In most cases,it may be wrong).Athough we can use cast() and convert() function to convert the data type,but it's better to speicify correct data type directly on account of consideration of performance.Notice that you'de better always keep functions in the right side.
 
Example
 
Test the implicit conversion in function convert() and cast().
 (zlm@192.168.1.101 )[zlm]>select cast('2abc' as unsigned) from dual;
+--------------------------+
| cast('2abc' as unsigned) |
+--------------------------+
| |
+--------------------------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[zlm]>select convert('2abc',unsigned) from dual;
+--------------------------+
| convert('2abc',unsigned) |
+--------------------------+
| |
+--------------------------+
row in set, warning (0.00 sec) //Function cast() and convert() turned the string '2abc' to number.
//The result became 2 what means the string begin with number only keep the first number remain. (zlm@192.168.1.101 )[zlm]>select convert('abc',unsigned) from dual;
+-------------------------+
| convert('abc',unsigned) |
+-------------------------+
| |
+-------------------------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[zlm]>select cast('abc' as unsigned) from dual;
+-------------------------+
| cast('abc' as unsigned) |
+-------------------------+
| |
+-------------------------+
row in set, warning (0.00 sec) //If there's no number prefix,the result turns out to be zero(integer type here). (zlm@192.168.1.101 )[zlm]>select ''=convert('3abc',unsigned);
+-----------------------------+
| ''=convert('abc',unsigned) |
+-----------------------------+
| |
+-----------------------------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[zlm]>select ''=cast('3abc' as unsigned);
+-----------------------------+
| ''=cast('abc' as unsigned) |
+-----------------------------+
| |
+-----------------------------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[zlm]>select =convert('3abc',unsigned);
+-----------------------------+
| ''=convert('abc',unsigned) |
+-----------------------------+
| |
+-----------------------------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[zlm]>select =cast('3abc' as unsigned);
+-----------------------------+
| ''=cast('abc' as unsigned) |
+-----------------------------+
| |
+-----------------------------+
row in set, warning (0.00 sec) //All the rusults are 1 what means the implicit conversion occurs.

Test the implicit conversion in expression and function concat().

 (zlm@192.168.1.101 )[zlm]>select +'abc' from dual;
+----------+
| +'abc' |
+----------+
| |
+----------+
row in set, warning (0.01 sec) //The string 'abc' converted implicitly from string to number and got result of 10(10 + 0 = 0). (zlm@192.168.1.101 )[zlm]>select concat(,'abc') from dual;
+------------------+
| concat(,'abc') |
+------------------+
| 10abc |
+------------------+
row in set (0.00 sec) //The number 10 was converted implicitly to string and concantenated by string 'abc'. (zlm@192.168.1.101 )[zlm]>select +'abc'=concat(,'abc');
+---------------------------+
| +'abc'=concat(,'abc') |
+---------------------------+
| |
+---------------------------+
row in set, warning (0.00 sec) //What may astonish us is that the result became 1,that is,implicit conversion occured again. (zlm@192.168.1.101 )[zlm]>select =concat(,'abc')-'abc';
+---------------------------+
| =concat(,'abc')-'abc' |
+---------------------------+
| |
+---------------------------+
row in set, warning (0.00 sec) //Moving the string 'abc' to the right side didn't influence the result of 1.

Test the implicit conversion in table.

 (zlm@192.168.1.101 )[sysbench]>desc sbtest1;
+-------+-----------+------+-----+---------+----------------+
| Field | Type | Null | Key | Default | Extra |
+-------+-----------+------+-----+---------+----------------+
| id | int() | NO | PRI | NULL | auto_increment |
| k | int() | NO | MUL | | |
| c | char() | NO | | | |
| pad | char() | NO | | | |
+-------+-----------+------+-----+---------+----------------+
rows in set (0.00 sec) (zlm@192.168.1.101 )[sysbench]>select * from sbtest1 limit ;
+----+------+-------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------+
| id | k | c | pad |
+----+------+-------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------+
| | | --------- | ---- |
| | | --------- | ---- |
| | | --------- | ---- |
| | | --------- | ---- |
| | | --------- | ---- |
+----+------+-------------------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------+
rows in set (0.00 sec) (zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where id='';
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| | SIMPLE | sbtest1 | NULL | const | PRIMARY | PRIMARY | | const | | 100.00 | NULL |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
row in set, warning (0.00 sec) (zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where id=cast( as char);
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
| | SIMPLE | sbtest1 | NULL | const | PRIMARY | PRIMARY | | const | | 100.00 | NULL |
+----+-------------+---------+------------+-------+---------------+---------+---------+-------+------+----------+-------+
row in set, warning (0.00 sec) //Despite I've specified 1 as string above,the queries still return result with primary key. (zlm@192.168.1.101 )[sysbench]>alter table sbtest1 add unique(c); //Add a unique key on column c(char type).
Query OK, rows affected ( min 58.75 sec) //It cost almost 5 mins.
Records: Duplicates: Warnings: (zlm@192.168.1.101 )[sysbench]>desc sbtest1;
+-------+-----------+------+-----+---------+----------------+
| Field | Type | Null | Key | Default | Extra |
+-------+-----------+------+-----+---------+----------------+
| id | int() | NO | PRI | NULL | auto_increment |
| k | int() | NO | MUL | | |
| c | char() | NO | UNI | | |
| pad | char() | NO | | | |
+-------+-----------+------+-----+---------+----------------+
rows in set (0.00 sec) (zlm@192.168.1.101 )[sysbench]>select count(*) from sbtest1;
+----------+
| count(*) |
+----------+
| |
+----------+
row in set (3.52 sec) //Query 1
(zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where c=---------;
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
| | SIMPLE | sbtest1 | NULL | ALL | c | NULL | NULL | NULL | | 10.00 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
row in set, warnings (0.00 sec) //Query 2
(zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where c='83868641912-28773972837-60736120486-75162659906-27563526494-20381887404-41576422241-93426793964-56405065102-33518432330';
+----+-------------+---------+------------+-------+---------------+------+---------+-------+------+----------+-------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+------+---------+-------+------+----------+-------+
| | SIMPLE | sbtest1 | NULL | const | c | c | | const | | 100.00 | NULL |
+----+-------------+---------+------------+-------+---------------+------+---------+-------+------+----------+-------+
row in set, warning (0.01 sec) //Query 3
(zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where c like ;
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
| | SIMPLE | sbtest1 | NULL | ALL | c | NULL | NULL | NULL | | 11.11 | Using where |
+----+-------------+---------+------------+------+---------------+------+---------+------+---------+----------+-------------+
row in set, warnings (0.00 sec) //Query 4
(zlm@192.168.1.101 )[sysbench]>explain select * from sbtest1 where c like '83868641912%';
+----+-------------+---------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+
| id | select_type | table | partitions | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+---------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+
| | SIMPLE | sbtest1 | NULL | range | c | c | | NULL | | 100.00 | Using index condition |
+----+-------------+---------+------------+-------+---------------+------+---------+------+------+----------+-----------------------+
row in set, warning (0.00 sec) //The implicit conversion occured in query 1 and query 3.They cannot use the index on column c.
//In query 4,it even used ICP feature of MySQL(Which is supported since version 5.6).
Summary
  • The implicit conversion usually occurs to make the query to be as compatible as possible in MySQL.
  • The implicit conversion occurs in expression,function and condiction of queries.
  • The implicit conversion may lead to bad performance because it will prevent MySQL from using indexes on specific query columns.
  • We'd better specify the correct data type explicit when typing them in our query SQL statement to avoid the implicit conversion.

MySQL隐式转换测试的更多相关文章

  1. MySQL隐式转换的坑

    MySQL以以下规则描述比较操作如何进行转换: 两个参数至少有一个是 NULL 时,比较的结果也是 NULL,例外是使用 <=> 对两个 NULL 做比较时会返回 1,这两种情况都不需要做 ...

  2. 一个 MySQL 隐式转换的坑,差点把服务器整崩溃了

    我是风筝,公众号「古时的风筝」,专注于 Java技术 及周边生态. 文章会收录在 JavaNewBee 中,更有 Java 后端知识图谱,从小白到大牛要走的路都在里面. 本来是一个平静而美好的下午,其 ...

  3. 关于MySQL隐式转换

    一.如果表定义的是varchar字段,传入的是数字,则会发生隐式转换. 1.表DDL 2.传int的sql 3.传字符串的sql 仔细看下表结构,rid的字段类型: 而用户传入的是int,这里会有一个 ...

  4. Mysql 隐式转换

    表定义: CREATE TABLE `ids` ( id ) not null auto_increment, PRIMARY KEY (id) ); 表中存在一些IDs: 111, 112, 113 ...

  5. Mysql隐式类型转换原则

    MySQL 的隐式类型转换原则: - 两个参数至少有一个是 NULL 时,比较的结果也是 NULL,例外是使用 <=> 对两个 NULL 做比较时会返回 1,这两种情况都不需要做类型转换 ...

  6. MySQL SQL优化之字符串索引隐式转换

    之前有用户很不解:SQL语句非常简单,就是select * from test_1 where user_id=1 这种类型,而且user_id上已经建立索引了,怎么还是查询很慢? test_1的表结 ...

  7. MySQL性能优化:MySQL中的隐式转换造成的索引失效

    数据库优化是一个任重而道远的任务,想要做优化必须深入理解数据库的各种特性.在开发过程中我们经常会遇到一些原因很简单但造成的后果却很严重的疑难杂症,这类问题往往还不容易定位,排查费时费力最后发现是一个很 ...

  8. mysql中的隐式转换

    在mysql查询中,当查询条件左右两侧类型不匹配的时候会发生隐式转换,可能导致查询无法使用索引.下面分析两种隐式转换的情况 看表结构 phone为 int类型,name为 varchar EXPLAI ...

  9. MySQL隐式转化整理

    MySQL隐式转化整理 前几天在微博上看到一篇文章:价值百万的 MySQL 的隐式类型转换感觉写的很不错,再加上自己之前也对MySQL的隐式转化这边并不是很清楚,所以就顺势整理了一下.希望对大家有所帮 ...

随机推荐

  1. MySQL入门很简单: 4 创建 修改删除表

    1. 创建表的方法 1)创建表的语法形式 首先,选择数据库: USE 数据库名: 创建表: CREATE TABLE 表名 (属性名 数据类型 [完整性约束条件], 属性名 数据类型 [完整性约束条件 ...

  2. 2017.10.28 针对Java Web应用中错误异常处理方法的运用

    针对Java Web应用中错误异常处理方法的运用 在javaweb中其异常都需要对Checked Exception之下的Exception进行继承,并且有选择地对发生的错误和异常进行处理.Java同 ...

  3. Linux---who命令学习

    who命令 获取正在登录系统的用户 使用Linux的who命令 第一个参数book代表用户名,第二个参数tty7代表终端名,第三个参数代表时间,第四个参数代表用户的登录地址. 阅读手册 使用命令读手册 ...

  4. 散度(Divergence)和旋度(Curl)

    原文链接 散度(Divergence) 散度的讨论应从向量和向量场说起.向量是数学中研究多维计算的基本概念.比如,速度可以分解为相互独立的分量,则速度就是一个多维的向量.假如空间中的每一个位置都有一个 ...

  5. data-ng-app 指令

    1.data-ng-app指令定义了一个AngularJS应用程序的根元素. 2.data-ng-app会在页面加载完毕后自动进行初始化应用程序. 3.data-ng-app可以通过一个值连接到代码模 ...

  6. this指向问题(1)

    在JS中,this一般有四种绑定的方式,但是在确定到底是哪种绑定之前必须先找到函数的调用位置.接下来先介绍其中的三种: 1.默认绑定 其实所谓的默认绑定就是函数直接调用(前面没有什么东西来点它),在默 ...

  7. OceanBase安装

    背景: OceanBase是阿里巴巴.蚂蚁金服自主研发的可扩展的分布式关系数据库,实现了数千亿条记录.数百 TB 数据上的跨行跨表事务,主要支持支付宝核心的交易.支付.会员和账务系统等 OLTP 和  ...

  8. HTML5一些标签和属性

    <bdo> 元素 可以覆盖默认文本的方向  根据dir 属性来控制文字的排序方向            属性:dir="rtl"                     ...

  9. Windows Server 2012 搭建DHCP及远程路由访问

    1.1    基础环境信息 1.2    DHCP与远程访问服务器角色安装 1.服务器管理器—>仪表板—>添加角色和功能,出现添加角色和功能向导,点击下一步 2.选择安装类型为基于角色或基 ...

  10. Drazil and Tiles CodeForces - 516B (类拓扑)

    Drazil created a following problem about putting 1 × 2 tiles into an n × m grid: "There is a gr ...