该表的功能是，防止用户注册过程中流逝，记录已经填写的数据。

			
CREATE TABLE `signup_keyloggers` (
	`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT COMMENT '唯一ID',
	`cookie` VARCHAR(32) NOT NULL COMMENT 'cookie id',
	`type` ENUM('baidu','google') NOT NULL COMMENT '推广账号类型',
	`field` ENUM('Name','Mobile','Email') NOT NULL COMMENT '字段名',
	`value` VARCHAR(50) NOT NULL COMMENT '值',
	`status` ENUM('New','Sent','Ignored','Called','Processed') NOT NULL DEFAULT 'New' COMMENT '状态',
	`operator` VARCHAR(10) NOT NULL COMMENT '操作人',
	`ctime` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',
	`mtime` TIMESTAMP NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '状态修改时间',
	PRIMARY KEY (`id`),
	UNIQUE INDEX `unique_index` (`type`, `cookie`, `field`, `value`)
)
COMMENT='用户注册键盘记录器'
COLLATE='utf8_general_ci'
ENGINE=InnoDB;
			
			

当用户注册成功会根据cookie id 删除该表中的数据。

当数据被记录后，客服就可以对客户回访，并修改状态status，忽略 Ignored，邮件发送Sent， 电话回访Called等等

			
 +-----------+
 | category  |
 |-----------|
 |id         | <---+
 |title      |     |
 |description|    1:n
 |status     |     |
 |parent_id  | o---+
 +-----------+
			
			

		
CREATE TABLE `category` (
	`id` SMALLINT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
	`name` VARCHAR(10) NOT NULL,
	`description` VARCHAR(255) NULL,
	`status` ENUM('enable','desable') NOT NULL DEFAULT 'enable',
	`parent_id` SMALLINT(10) UNSIGNED NOT NULL DEFAULT '0',
	PRIMARY KEY (`id`),
	CONSTRAINT `FK1` FOREIGN KEY (`parent_id`) REFERENCES `category` (`id`)
)
COMMENT='goods category'
ENGINE=InnoDB
ROW_FORMAT=DEFAULT
		
			

多对多分类,主要用于满足，一个产品/文章属于多个分类的需求。

		
      +------------+
      | category   |
      |------------|
 +--> |id          | <---+
 |    |title       |     |     +----------------------+
1:n   |description |    1:n    | categroy_has_product |
 |    |status      |     |     +----------------------+
 +--o |parent_id   |     |     | id                   |
      +------------+     +---o | category_id          |
                         +---o | product_id           |
      +------------+     |     +----------------------+
      | product    |    1:n
      +------------+     |
      |id          | <---+
      |price       |
      |quantity    |
      |...         |
      |status      |
      +------------+
		
			

上面我刚刚讲过怎样实现“不限子树的分类树”，我们可以实现不限层次的无线分类表。

			
 +-----------+
 | category  |
 |-----------|
 |id         | <---+
 |title      |     |
 |description|    1:n
 |status     |     |
 |parent_id  | o---+
 +-----------+
			
			

问题出来了，当我需要读取一个分类（任意分类）下的所有子分类，怎样实现，很多人会说用“递归”。 当然“递归”可是现实我们的需求，在几百个分类的项目中，使用递归也不是不可以的，但是当数量非常庞大时怎么办？

当然有更好的解决方案，请看下面

			
 +-----------+
 | category  |
 |-----------|
 |id         | <---+
 |title      |     |
 |description|    1:n
 |status     |     |
 |parent_id  | o---+
 |path       |
 +-----------+
			
			

			
+-------------------------------------------------------------------------+
| category                                                                |
+----+-----------+-----------------------+--------+-----------+-----------+
| id | name      | description           | status | parent_id | path      |
+----+-----------+-----------------------+--------+-----------+-----------+
|  1 | 中国    | 中华人民共和家                                    | Y      |      NULL | 1/        |
|  4 | 广东省 | 广东省                                                      | Y      |         1 | 1/4       |
|  5 | 深圳市 | NULL                      | Y      |         4 | 1/4/5     |
|  6 | 宝安区 | NULL                      | Y      |         5 | 1/4/5/6   |
|  7 | 龙华镇 | NULL                      | Y      |         6 | 1/4/5/6/7 |
+----+-----------+-----------------------+--------+-----------+-----------+
			
			

			
CREATE TABLE `category` (
	`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT COMMENT '分类ID',
	`name` VARCHAR(50) NOT NULL COMMENT '分类名称',
	`description` VARCHAR(200) NULL DEFAULT NULL COMMENT '分类描述',
	`status` ENUM('Y','N') NOT NULL DEFAULT 'Y' COMMENT '分类状态有继承性',
	`parent_id` INT(10) NULL DEFAULT '1' COMMENT '分类父ID',
	`path` VARCHAR(255) NOT NULL COMMENT '分类递归路径索引',
	INDEX `PK` (`id`),
	INDEX `relation` (`id`, `parent_id`),
	INDEX `FK_category_category` (`parent_id`),
	INDEX `path` (`path`)
)
COMMENT='分类表'
ENGINE=InnoDB
ROW_FORMAT=DEFAULT
AUTO_INCREMENT=0

insert into category(`name`,`description`,`status`,`parent_id`,`path`) values('中国','中华人民共和家','Y',null,'1/')
			
			

			
ALTER TABLE `category`
	ADD CONSTRAINT `FK_category_category` FOREIGN KEY (`parent_id`) REFERENCES `category` (`id`)
			
			

抽取广东子树

			
select * from category where path like '1/4%';
			
			

			
mysql> select * from category where path like '1/4%';
+----+-----------+-------------+--------+-----------+-----------+
| id | name      | description | status | parent_id | path      |
+----+-----------+-------------+--------+-----------+-----------+
|  4 | 广东省 | 广东省   | Y      |         1 | 1/4       |
|  5 | 深圳市 | NULL        | Y      |         4 | 1/4/5     |
|  6 | 宝安区 | NULL        | Y      |         5 | 1/4/5/6   |
|  7 | 龙华镇 | NULL        | Y      |         6 | 1/4/5/6/7 |
+----+-----------+-------------+--------+-----------+-----------+
4 rows in set (0.00 sec)
			
			

			
DROP TABLE IF EXISTS `test`;
CREATE TABLE IF NOT EXISTS `test` (
  `id` int(11) DEFAULT NULL,
  `pid` int(11) DEFAULT NULL,
  `name` char(50) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

INSERT INTO `test` (`id`, `pid`, `name`) VALUES
	(1, 0, 'A'),
	(2, 1, 'B'),
	(3, 1, 'C'),
	(4, 0, 'D'),
	(5, 0, 'E'),
	(6, 5, 'F');

select (select t2.name from test t2 where t2.id=t1.pid) as name, count(pid) as sum from test t1 where t1.pid <> 0 group by t1.pid;
			
			

统计所有节点包括数量为零的

				select t1.name, (select count(t2.name) from test t2 where t2.pid=t1.id) as sum from test t1
			

分区表可以通过表空间，等等技术实现，优点是解决了Union查询问题，保证了数据的一致性。

		
      +-----------+
      | category  |
      |-----------|
  +-->|id         | <---+
  |   |title      |     |
  |   |description|    1:n
  |   |status     |     |
  |   |parent_id  | o---+
  |   +-----------+
  |
 1:n
  |
  |   +-----------------+            +-----------------+
  |   | article         |            | feedback        |
  |   |-----------------|            |-----------------|
  |   |id               |<--1:n--+   |id               |
  |   |title            |        |   |title            |
  |   |content          |        |   |content          |
  |   |datetime         |        |   |datetime         |
  |   |status           |        |   |status           |
  +--o|category_id      |        +--o|news_id          |
  +--o|user_id          |        +-->|user_id          |
  |   +-----------------+        |   +-----------------+
  |   | 2007,2008,2009  |        |   | 2007,2008,2009  |
  |   +-----------------+        |   +-----------------+
  |                              |
 1:n  +----------+     +---1:n---+
  |   | user     |     |
  |   |----------|     |
  +-->|id        | <---+
      |user      |
      |passwd    |
      |nickname  |
      |status    |
      +----------+
		
			

显示title前20个汉字并在后尾添加省略号。

		
+------------+           +--------------------------+           +-----------------------+
| product    |           | product_attribute        |           |product_attribute_key  |
+------------+           +--------------------------+           +-----------------------+
|id          | <--1:1--o |product_id                |     +---> |id                     |
|price       |           |product_attribute_key_id  | o---+     |name                   |
|quantity    |           |product_attribute_value_id| o---+     +-----------------------+
|...         |           +--------------------------+     |     +-----------------------+
|category_id |                                           1:n    |product_attribute_value|
+------------+                                            |     +-----------------------+
                                                          +---> |id                     |
                                                                |name                   |
                                                                +-----------------------+
		
			

product attribute group

		
+------------+           +--------------------------+           +--------------------------+           +-----------------------+
| category   |           | product_attribute_group  |           | product_attribute        |           |product_attribute_key  |
+------------+           +--------------------------+           +--------------------------+           +-----------------------+
|id          |     +---> |id                        | <--1:n--o |product_attribute_group_id|     +---> |id                     |
|title       |     |     |name                      |           |product_attribute_key_id  | o---+     |name                   |
|description |    1:1    |status                    |           |product_attribute_value_id| o---+     +-----------------------+
|status      |     |     +--------------------------+           +--------------------------+     |     +-----------------------+
|parent_id   |     |                                                                            1:n    |product_attribute_value|
|default_pag | o---+                                                                             |     +-----------------------+
+------------+                                                                                   +---> |id                     |
                                                                                                       |name                   |
                                                                                                       +-----------------------+
		
			

product attribute group

		    
    +------------+           +------------------+           +--------------------------+           +---------------------------------+
    | category   |           | attribute_group  |           | group_has_attribute      |           |attribute_key                    |
    +------------+           +------------------+           +--------------------------+           +---------------------------------+
 +->|id          |     +-->  |id                | <--1:n--o |attribute_group_id        |     +-+-> |id                               |
 |  |title       |     |     |name              |           |attribute_key_id          | o---+ |   |name                             |
 |  |description |    1:1    |status            |           |                          |       |   |type enum('Bool','List','Input') |
 |  |status      |     |     +------------------+           +--------------------------+       |   |default array()                  |
 |  |parent_id   |     |                                                                       |   +---------------------------------+
 |  |default_pag | o---+                                                                       |
 |  +------------+                                                                             |
1:n                                                                                            |
 |  +-------------+             +--------------------------+                                   |
 |  | product     |             | product_attribute        |                                   |
 |  +-------------+             +--------------------------+                                   |
 |  |id           |         +-> |product_id                |                                   |
 |  |price        |         |   |attribute_key_id          | o---------------1:n---------------+
 |  |quantity     |         |   |attribute_value           |
 |  |...          |         |   +--------------------------+
 +-o|category_id  |         |
    |attr_group_id| <--1:n--o
    +-------------+
		    
			

				+--------------------------------------------------+
				| product_attribute_key |
				+--------------------------------------------------+
				| 1 | color | list | red,green,blue |
				| 2 | sex | bool | Female,Male |
				| 3 | qty | input| '' |
				+--------------------------------------------------+
			

我有一个表，里面只有固定行数的行记录，这些数据就是配置参数，我们将配置文件保存在数据库中，因为需要做负载均衡而不能使用文件配置文件。

有这样一个需求，这个记录每次修改都要保存历史记录，用于审计等等。我是这样设计该表的

				CREATE TABLE `config_fee` (
				`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
				`level` INT(11) NULL DEFAULT NULL COMMENT '层级',
				`type` ENUM('Deposit','Withdrawing') NOT NULL DEFAULT 'Withdrawing' COMMENT '类型，存款，取款',
				`min_fee` FLOAT(10,2) NOT NULL COMMENT '最低手续费',
				`max_fee` FLOAT(10,2) NOT NULL COMMENT '最高手续费',
				`ratio` FLOAT(10,2) NOT NULL COMMENT '手续费比例',
				`operator` VARCHAR(10) NOT NULL COMMENT '操作者',
				`status` ENUM('Current','Trash') NOT NULL DEFAULT 'Current',
				`ctime` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
				`mtime` TIMESTAMP NULL DEFAULT NULL ON UPDATE CURRENT_TIMESTAMP,
				PRIMARY KEY (`id`)
				)
				COMMENT='手续费管理'
				COLLATE='utf8_general_ci'
				ENGINE=InnoDB;
			

数据记录的形态

			
mysql> select type,operator,status,ctime,mtime from config_mtf_fee;
+---------+----------+---------+---------------------+---------------------+
| type    | operator | status  | ctime               | mtime               |
+---------+----------+---------+---------------------+---------------------+
| Deposit | 141      | Trash   | 2014-08-28 11:10:17 | 2014-08-28 11:10:57 |
| Deposit | 141      | Trash   | 2014-08-28 11:10:17 | 2014-08-28 11:10:57 |
| Deposit | 141      | Trash   | 2014-08-28 11:10:17 | 2014-08-28 11:10:57 |
| Deposit | 141      | Trash   | 2014-08-28 11:10:17 | 2014-08-28 11:10:57 |
| Deposit | 324      | Current | 2014-08-28 11:10:54 | 2014-08-28 11:10:59 |
+---------+----------+---------+---------------------+---------------------+
2 rows in set (0.00 sec)
			
			

如上图所示，状态 Current 是当前记录，而Trash是废弃的历史记录。

每次修改数据，首先将Current改为Trash，然后插入一条新数据状态为Current，我们只会使用最后一条状态为current的数据。

该函数能够检查身份证号码是否正确

			
CREATE DEFINER=`neo`@`%` FUNCTION `check_id_number`(`idnumber` CHAR(18))
	RETURNS enum('true','false')
	LANGUAGE SQL
	NOT DETERMINISTIC
	NO SQL
	SQL SECURITY DEFINER
	COMMENT ''
BEGIN
DECLARE status ENUM('true','false') default 'false';
DECLARE verify CHAR(1);
DECLARE sigma INT;
DECLARE remainder INT;

IF length(idnumber) = 18 THEN
	set sigma = cast(substring(idnumber,1,1) as UNSIGNED) * 7
		+cast(substring(idnumber,2,1) as UNSIGNED) * 9
		+cast(substring(idnumber,3,1) as UNSIGNED) * 10
		+cast(substring(idnumber,4,1) as UNSIGNED) * 5
		+cast(substring(idnumber,5,1) as UNSIGNED) * 8
		+cast(substring(idnumber,6,1) as UNSIGNED) * 4
		+cast(substring(idnumber,7,1) as UNSIGNED) * 2
		+cast(substring(idnumber,8,1) as UNSIGNED) * 1
		+cast(substring(idnumber,9,1) as UNSIGNED) * 6
		+cast(substring(idnumber,10,1) as UNSIGNED) * 3
		+cast(substring(idnumber,11,1) as UNSIGNED) * 7
		+cast(substring(idnumber,12,1) as UNSIGNED) * 9
		+cast(substring(idnumber,13,1) as UNSIGNED) * 10
		+cast(substring(idnumber,14,1) as UNSIGNED) * 5
		+cast(substring(idnumber,15,1) as UNSIGNED) * 8
		+cast(substring(idnumber,16,1) as UNSIGNED) * 4
		+cast(substring(idnumber,17,1) as UNSIGNED) * 2;
	set remainder = MOD(sigma,11);
	set verify = (case remainder
		when 0 then '1' when 1 then '0' when 2 then 'X' when 3 then '9'
		when 4 then '8' when 5 then '7' when 6 then '6' when 7 then '5'
		when 8 then '4' when 9 then '3' when 10 then '2' else '/' end
	);

END IF;

IF right(idnumber,1) = verify THEN
	set status = 'true';
END IF;

RETURN status;

END
			
			

首先我们使用正确身份证号码进行测试，返回true

			
mysql> select check_id_number('330702198003090915');
+---------------------------------------+
| check_id_number('330702198003090915') |
+---------------------------------------+
| true                                  |
+---------------------------------------+
1 row in set (0.01 sec)
			
			

长度不符合18位直接返回false.

			
mysql> select check_id_number('33070219800309');
+-----------------------------------+
| check_id_number('33070219800309') |
+-----------------------------------+
| false                             |
+-----------------------------------+
1 row in set (0.00 sec)	

mysql> select check_id_number('33070219800309091457889');
+--------------------------------------------+
| check_id_number('33070219800309091457889') |
+--------------------------------------------+
| false                                      |
+--------------------------------------------+
1 row in set, 1 warning (0.00 sec)
			
			

随便改译为数，校验失败返回 false

			
mysql> select check_id_number('330702198003090914');
+---------------------------------------+
| check_id_number('330702198003090914') |
+---------------------------------------+
| false                                 |
+---------------------------------------+
1 row in set (0.00 sec)					
			
			

people 是存储人，你可以用为user,member都可以

			
CREATE TABLE `people` (
	`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
	`name` VARCHAR(50) NOT NULL,
	PRIMARY KEY (`id`)
)
COMMENT='Social Network Site - Six Degrees of Separation - http://www.netkiller.cn'
COLLATE='utf8_general_ci'
ENGINE=InnoDB;
			
			

存储具体的这人

这个表的功能主要是维持朋友之间的关系网，这里使用了多对多方式并且使用外键防止产生脏数据。

			
CREATE TABLE `friend` (
	`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
	`people_id` INT(10) UNSIGNED NOT NULL,
	`friend_id` INT(10) UNSIGNED NOT NULL,
	`ctime` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
	PRIMARY KEY (`id`),
	UNIQUE INDEX `unique` (`people_id`, `friend_id`),
	INDEX `FK_firend_people` (`people_id`),
	INDEX `FK_firend_people_2` (`friend_id`),
	CONSTRAINT `FK_firend_people` FOREIGN KEY (`people_id`) REFERENCES `people` (`id`),
	CONSTRAINT `FK_firend_people_2` FOREIGN KEY (`friend_id`) REFERENCES `people` (`id`)
)
COMMENT='Social Network Site - Six Degrees of Separation - http://www.netkiller.cn'
COLLATE='utf8_general_ci'
ENGINE=InnoDB;

			
			

首先初始化用户数据

			
INSERT INTO `people` (`id`, `name`) VALUES
	(1, 'Neo'),
	(2, 'Luke'),
	(3, 'Jack'),
	(4, 'Joey'),
	(5, 'Jam'),
	(6, 'John');
			
			

建立朋友之间的关系

			
INSERT INTO `friend` (`id`, `people_id`, `friend_id`) VALUES
	(1, 1, 2),
	(2, 1, 3),
	(3, 1, 4),
	(4, 1, 5),
	(5, 1, 6),
	(6, 2, 1),
	(7, 2, 3);
			
			

现在就可以查找你的朋友了

			
select people.* from friend, people where friend.people_id = 1 and friend.friend_id = people.id;
			
			

查找朋友的朋友就比较麻烦了，必须使用递归方法，一层一层查下去，反复执行SQL效率是很低的，所以我们准备了第三张表。

关系网表，主要功能是弥补firend表，用于快速检索（在不使用递归的情况下）

			
CREATE TABLE `network` (
	`id` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
	`people_id` INT(10) UNSIGNED NOT NULL,
	`following_id` INT(10) UNSIGNED NOT NULL,
	`friend_id` INT(10) UNSIGNED NULL DEFAULT NULL,
	`degrees` VARCHAR(250) NOT NULL,
	`ctime` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP,
	PRIMARY KEY (`id`),
	UNIQUE INDEX `unique` (`people_id`, `friend_id`, `following_id`),
	INDEX `FK_firend_people` (`people_id`),
	INDEX `FK_firend_people_2` (`friend_id`),
	INDEX `FK_friend_people_following_id` (`following_id`),
	CONSTRAINT `FK_firend_people` FOREIGN KEY (`people_id`) REFERENCES `people` (`id`),
	CONSTRAINT `FK_friend_people_following_id` FOREIGN KEY (`following_id`) REFERENCES `people` (`id`),
	CONSTRAINT `FK_friend_people_friend_id` FOREIGN KEY (`friend_id`) REFERENCES `people` (`id`)
)
COMMENT='Social Network Site - Six Degrees of Separation - http://www.netkiller.cn'
COLLATE='utf8_general_ci'
ENGINE=InnoDB;

			
			

following 一个朋友, Neo following Jam

			
INSERT INTO `people` (`id`, `name`) VALUES
	(1, 'Neo'),
	(2, 'Luke'),
	(3, 'Jack'),
	(4, 'Joey'),
	(5, 'Jam'),
	(6, 'John');

INSERT INTO `network` (`people_id`, `following_id`, `friend_id`, `degrees`) VALUES ( 1, 5, NULL, '1.5');
			
			

之前Neo已经 following Jam，接下来查找Jam的朋友，现在Neo following John, John 是 Jam 的朋友，friend_id = NULL 表示 Jam 尚未有朋友

			
select * from network where people_id=1 and friend_id = 5;

INSERT INTO `sns`.`network` (`people_id`, `following_id`, `friend_id`, `degrees`) VALUES ('1', '6', '5', '1.5.6');
			
			

Neo following Joey, Joey 是 Luke 的朋友， 所以 Luke可能是 Neo的朋友

			
INSERT INTO `sns`.`network` (`people_id`, `following_id`, `friend_id`, `degrees`) VALUES ('1', '4', '2', '1.2.4');
			
			

查询不同维度下的所有好友，查询出的用户ID需要处理。

			
select * from network where people_id=1 and degrees like "1.%";
select * from network where people_id=1 and degrees like "1.2%";
select * from network where people_id=1 and degrees like "1.2.%";
			
			

至此社区管理网就建立起来了

上面的例子演示了 people_id=1 即 Neo 的关系网

			
 +-----------+
 | city      |
 |-----------|
 |id         | <---+
 |name       |     |
 |description|    1:n
 |status     |     |
 |parent_id  | o---+
 +-----------+
			
			

	    	
CREATE TABLE city
(
  id serial NOT NULL,
  name character varying,
  parent_id integer,
  status boolean,
  CONSTRAINT city_pkey PRIMARY KEY (id),
  CONSTRAINT city_parent_id_fkey FOREIGN KEY (parent_id)
      REFERENCES city (id) MATCH SIMPLE
      ON UPDATE NO ACTION ON DELETE NO ACTION
)
WITH (
  OIDS=FALSE
);
ALTER TABLE city
  OWNER TO sys;
	    	
	    	

	    	
INSERT INTO city (id, name, parent_id, status) VALUES (1, '广东', NULL, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (2, '湖南', NULL, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (3, '深圳', 1, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (4, '东莞', 1, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (5, '福田', 3, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (6, '南山', 3, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (7, '宝安', 3, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (8, '西乡', 7, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (9, '福永', 7, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (10, '龙华', 7, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (11, '长沙', 2, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (12, '湘潭', 2, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (13, '常德', 2, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (14, '桃源', 13, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (15, '汉寿', 13, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (16, '黑龙江', NULL, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (17, '伊春', 16, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (18, '哈尔滨', 16, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (19, '齐齐哈尔', 16, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (20, '牡丹江', 16, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (21, '佳木斯', 16, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (22, '民治', 10, NULL);
INSERT INTO city (id, name, parent_id, status) VALUES (23, '上塘', 10, NULL);
	    	
	    	

	    	
WITH RECURSIVE path(id, name, path, idpath, parent_id, status) AS (
  SELECT id, name, '/' || name , '/' || id , parent_id, status FROM city WHERE parent_id is null
  UNION
  SELECT
    city.id,
    city.name,
    parentpath.path ||
      CASE parentpath.path
	WHEN '/' THEN ''
	ELSE '/'
      END || city.name,
    parentpath.idpath ||
     CASE parentpath.idpath
	WHEN '/' THEN ''
	ELSE '/'
      END || city.id,
    city.parent_id, city.status
  FROM city, path as parentpath
  WHERE city.parent_id = parentpath.id
)

SELECT * FROM path;
	    	
	    	

	    	
 id |   name   |           path            |    idpath    | parent_id | status
----+----------+---------------------------+--------------+-----------+--------
  1 | 广东     | /广东                     | /1           |           |
  2 | 湖南     | /湖南                     | /2           |           |
 16 | 黑龙江   | /黑龙江                   | /16          |           |
  3 | 深圳     | /广东/深圳                | /1/3         |         1 |
  4 | 东莞     | /广东/东莞                | /1/4         |         1 |
 11 | 长沙     | /湖南/长沙                | /2/11        |         2 |
 12 | 湘潭     | /湖南/湘潭                | /2/12        |         2 |
 13 | 常德     | /湖南/常德                | /2/13        |         2 |
 17 | 伊春     | /黑龙江/伊春              | /16/17       |        16 |
 18 | 哈尔滨   | /黑龙江/哈尔滨            | /16/18       |        16 |
 19 | 齐齐哈尔 | /黑龙江/齐齐哈尔          | /16/19       |        16 |
 20 | 牡丹江   | /黑龙江/牡丹江            | /16/20       |        16 |
 21 | 佳木斯   | /黑龙江/佳木斯            | /16/21       |        16 |
  5 | 福田     | /广东/深圳/福田           | /1/3/5       |         3 |
  6 | 南山     | /广东/深圳/南山           | /1/3/6       |         3 |
  7 | 宝安     | /广东/深圳/宝安           | /1/3/7       |         3 |
 14 | 桃源     | /湖南/常德/桃源           | /2/13/14     |        13 |
 15 | 汉寿     | /湖南/常德/汉寿           | /2/13/15     |        13 |
  8 | 西乡     | /广东/深圳/宝安/西乡      | /1/3/7/8     |         7 |
  9 | 福永     | /广东/深圳/宝安/福永      | /1/3/7/9     |         7 |
 10 | 龙华     | /广东/深圳/宝安/龙华      | /1/3/7/10    |         7 |
 22 | 民治     | /广东/深圳/宝安/龙华/民治 | /1/3/7/10/22 |        10 |
 23 | 上塘     | /广东/深圳/宝安/龙华/上塘 | /1/3/7/10/23 |        10 |
(23 rows)
	    	
	    	

此表适用于购物车等金钱来往账面等等。

		
-- Table: account

-- DROP TABLE account;

CREATE TABLE account
(
  id integer NOT NULL DEFAULT nextval('trade_id_seq'::regclass),
  no character varying(10) NOT NULL, -- 账号
  balance money NOT NULL DEFAULT 0.00, -- 余额
  datetime timestamp without time zone NOT NULL DEFAULT (now())::timestamp(0) without time zone,
  CONSTRAINT account_pkey PRIMARY KEY (id)
)
WITH (
  OIDS=FALSE
);
ALTER TABLE account
  OWNER TO dba;
COMMENT ON COLUMN account.no IS '账号';
COMMENT ON COLUMN account.balance IS '余额';


-- Index: account_no_idx

-- DROP INDEX account_no_idx;

CREATE INDEX account_no_idx
  ON account
  USING btree
  (no COLLATE pg_catalog."default");
		
		
		

账户结余计算

		
select acc.*, (select sum(balance)+acc.balance from account as ac where ac.id < acc.id) as profit from account as acc;

test=# select acc.*, (select sum(balance)+acc.balance from account as ac where ac.id < acc.id) as profit from account as acc;
 id |  no  | balance  |      datetime       | profit
----+------+----------+---------------------+---------
  1 | 1000 |    $0.00 | 2013-10-09 10:51:10 |
  2 | 1000 |   $12.60 | 2013-10-09 10:51:22 |  $12.60
  4 | 1000 |   $16.80 | 2013-10-09 10:51:42 |  $29.40
  5 | 1000 |  $100.00 | 2013-10-09 10:51:49 | $129.40
  6 | 1000 |  $200.00 | 2013-10-09 10:56:35 | $329.40
  7 | 1000 |   $50.45 | 2013-10-09 10:57:23 | $379.85
  8 | 1000 |   $75.50 | 2013-10-09 10:57:31 | $455.35
  9 | 1000 |  -$55.30 | 2013-10-09 10:59:28 | $400.05
 10 | 1000 | -$200.00 | 2013-10-09 10:59:44 | $200.05
(9 rows)
				
		

		
ALTER TABLE `goods`  DROP INDEX `goods_sn_2`;
ALTER TABLE goods PARTITION BY RANGE (goods_id) (
    PARTITION p0 VALUES LESS THAN (10000),
    PARTITION p1 VALUES LESS THAN (20000),
    PARTITION p2 VALUES LESS THAN (30000),
    PARTITION p3 VALUES LESS THAN (40000),
    PARTITION p4 VALUES LESS THAN MAXVALUE
);

ALTER TABLE goods PARTITION BY HASH(goods_id) PARTITIONS 10;

ALTER TABLE goods  PARTITION BY KEY (is_on_sale) PARTITIONS 2;

ALTER TABLE goods PARTITION BY HASH(YEAR(FROM_UNIXTIME(add_time))) PARTITIONS 4;
		
		

这里我通过一个新闻网站为例，解决分表的问题

避免开发中经常拼接表，我采用一个一劳永逸的方法，建立一个 news 表使用黑洞引擎，然后通过出发器将数据分流到匹配的表中。同时采用uuid替代数字序列，可以保证未来数年不会出现ID用尽。

		
CREATE TABLE IF NOT EXISTS `news` (
  `uuid` varchar(36) NOT NULL COMMENT '唯一ID',
  `title` varchar(50) NOT NULL COMMENT '新闻标题',
  `body` text NOT NULL COMMENT '新闻正文',
  `ctime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '创建时间',
  `mtime` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '修改时间',
  `atime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '访问时间',
  PRIMARY KEY (`uuid`)
) ENGINE=BLACKHOLE DEFAULT CHARSET=utf8;
		
		

该表仅仅用于举例，结构比较简单。接下来创建年份分表，你也可以每个月一个表，根据你的许下灵活调整。表结构与上面的news表相同，注意 ENGINE=InnoDB。

		
CREATE TABLE IF NOT EXISTS `news_2012` (
  `uuid` varchar(36) NOT NULL COMMENT '唯一ID',
  `title` varchar(50) NOT NULL COMMENT '新闻标题',
  `body` text NOT NULL COMMENT '新闻正文',
  `ctime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '创建时间',
  `mtime` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '修改时间',
  `atime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '访问时间',
  PRIMARY KEY (`uuid`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COMMENT='news 表';

CREATE TABLE IF NOT EXISTS `news_2013` (
  `uuid` varchar(36) NOT NULL COMMENT '唯一ID',
  `title` varchar(50) NOT NULL COMMENT '新闻标题',
  `body` text NOT NULL COMMENT '新闻正文',
  `ctime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '创建时间',
  `mtime` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '修改时间',
  `atime` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' COMMENT '访问时间',
  PRIMARY KEY (`uuid`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COMMENT='news 表';
		
		

uuid 索引表，主要的功能是通过uuid查询出该记录在那张表中。更好的方案是将数据放入solr中处理，包括标题与内容搜索等等。

		
CREATE TABLE `news_index` (
	`uuid` VARCHAR(36) NOT NULL,
	`tbl_name` VARCHAR(10) NOT NULL,
	PRIMARY KEY (`uuid`)
)
COMMENT='news uuid 索引表'
COLLATE='utf8_general_ci'
ENGINE=InnoDB;

		
		

news_insert 过程，用于向目标表中插入数据，可以单独call 但不建议。因为insert 远比 call 更通用，要考虑移植性与通用性

		
DELIMITER //
CREATE DEFINER=`neo`@`%` PROCEDURE `news_insert`(IN `uuid` vARCHAR(36), IN `title` VARCHAR(50), IN `body` TEXT, IN `ctime` TIMESTAMP)
BEGIN
	if year(ctime) = '2012' then
		insert into news_2012(uuid,title,body,ctime) values(uuid,title, body, ctime);
	end if;
	if year(ctime) = '2013' then
		insert into news_2013(uuid,title,body,ctime) values(uuid,title, body, ctime);
	end if;
	insert into news_index values(uuid, year(ctime));
END//
DELIMITER ;
		
		

插入触发器，负责获取 uuid 然后调用存储过程

		
SET @OLDTMP_SQL_MODE=@@SQL_MODE, SQL_MODE='';
DELIMITER //
CREATE TRIGGER `news_before_insert` BEFORE INSERT ON `news` FOR EACH ROW BEGIN
	IF new.uuid is null or new.uuid = '' or length(new.uuid) != 36 THEN
		set new.uuid=uuid();
	END IF;
	call news_insert(new.uuid,new.title,new.body,new.ctime);
END//
DELIMITER ;
SET SQL_MODE=@OLDTMP_SQL_MODE;
		
		

这个触发器用户保护表中的 uuid 值不被修改。

		
SET @OLDTMP_SQL_MODE=@@SQL_MODE, SQL_MODE='';
DELIMITER //
CREATE TRIGGER `news_before_update` BEFORE UPDATE ON `news_2013` FOR EACH ROW BEGIN
	set new.uuid = old.uuid;
END//
DELIMITER ;
SET SQL_MODE=@OLDTMP_SQL_MODE;
		
		

delete from mytable 必死无疑，你需要分批删除，尽量缩小每个批次删除的记录数，delete 是可以并行执行的，你可以同时运行多个删除操作

		
mysql> show processlist;
+--------+-----------------+---------------------+-----------+---------+-------+-----------------------------+--------------------------------------------------------+
| Id     | User            | Host                | db        | Command | Time  | State                       | Info                                                   |
+--------+-----------------+---------------------+-----------+---------+-------+-----------------------------+--------------------------------------------------------+
|      1 | event_scheduler | localhost           | NULL      | Daemon  |    52 | Waiting for next activation | NULL                                                   |
| 115986 | dba             | localhost           | example   | Query   |     0 | NULL                        | show processlist                                       |
| 117446 | dba             | localhost           | example   | Query   |    20 | updating                    | delete from mytable where OPEN_TIME like '2011.11.28%' |
| 117525 | dba             | localhost           | example   | Query   |     2 | updating                    | delete from mytable where OPEN_TIME like '2011.12.02%' |
| 117526 | dba             | localhost           | example   | Query   |    49 | updating                    | delete from mytable where OPEN_TIME like '2011.12.12%' |
| 117527 | dba             | localhost           | example   | Query   |     6 | updating                    | delete from mytable where OPEN_TIME like '2011.12.21%' |
| 117528 | dba             | localhost           | example   | Query   |    64 | updating                    | delete from mytable where OPEN_TIME like '2011.12.30%' |
| 117546 | dba             | localhost           | example   | Query   |    33 | updating                    | delete from mytable where OPEN_TIME like '2011.11.10%' |
+--------+-----------------+---------------------+-----------+---------+-------+-----------------------------+--------------------------------------------------------+
23 rows in set (0.00 sec)
		
			

		
set global event_scheduler = on;		
		
			

			
备份 SQL：
CREATE TABLE mytable_2022_7_30 SELECT * FROM mytable;

删除 SQL：
DELETE FROM mytable LIMIT 1; 

回撤 SQL：
INSERT INTO mytable SELECT * FROM mytable_2022_7_30;	
			
			

			
备份 SQL：
CREATE TABLE mytable_2022_7_30 SELECT * FROM mytable where level = 'info';

删除 SQL：
DELETE FROM mytable where level = 'info' LIMIT 1; 

回撤 SQL：
INSERT INTO mytable SELECT * FROM mytable_2022_7_30;
			
			

			
DELIMITER $$
CREATE EVENT IF NOT EXISTS event_delete
ON SCHEDULE EVERY 1 SECOND ON COMPLETION PRESERVE 
DO BEGIN
	DECLARE num integer;

	SELECT  COUNT(*) INTO num FROM mytable;
    
	IF num > 0 THEN
		delete from mytable limit 1;
		insert into logs(ctime) values(now());
	END IF;
    
END$$
			
			

			
DROP EVENT event_delete;
			
			

在电商领域常常遇到一个问题“调价”，经常需要调整一批商品的价格, 程序猿一条语句搞定有没有？

		
update goods set price=price+10 where category_id = xxx
		
		

在开发，测试环境是可以通过测试的，一旦部署到生产环境，必死无疑

大表创建索引需要很久的时间，通常要经历 manage keys 与 copy to tmp table 的过程

		
mysql> show processlist;
+--------+-----------------+---------------------+----------+---------+-------+-----------------------------+------------------------------------------------------------------+
| Id     | User            | Host                | db       | Command | Time  | State                       | Info                                                             |
+--------+-----------------+---------------------+----------+---------+-------+-----------------------------+------------------------------------------------------------------+
|      1 | event_scheduler | localhost           | NULL     | Daemon  |    47 | Waiting for next activation | NULL                                                             |
| 115986 | dba             | localhost           | example  | Query   |     0 | NULL                        | show processlist                                                 |
| 118814 | dba             | 192.168.6.20:50459  | example  | Query   |     8 | copy to tmp table           | ALTER TABLE `mytable` ADD INDEX `modifiy_time` (`MODIFY_TIME`)   |
+--------+-----------------+---------------------+----------+---------+-------+-----------------------------+------------------------------------------------------------------+
17 rows in set (0.00 sec)
		
		

删除索引，也需要经理 copy to tmp table 过程，漫长的等待

		
mysql> show processlist;
+--------+-----------------+---------------------+--------------+---------+-------+-----------------------------+-------------------------------------------------+
| Id     | User            | Host                | db           | Command | Time  | State                       | Info                                            |
+--------+-----------------+---------------------+--------------+---------+-------+-----------------------------+-------------------------------------------------+
|      1 | event_scheduler | localhost           | NULL         | Daemon  |    11 | Waiting for next activation | NULL                                            |
| 115986 | dba             | localhost           | example      | Query   |     0 | NULL                        | show processlist                                |
| 118814 | dba             | 192.168.6.20:50459  | example      | Query   |     4 | copy to tmp table           | ALTER TABLE `mytable`	DROP INDEX `modifiy_time` |
+--------+-----------------+---------------------+--------------+---------+-------+-----------------------------+-------------------------------------------------+
17 rows in set (0.00 sec)
		
		

所以数据设计要深思熟虑，做到提前未雨绸缪，不要亡羊补牢

OPTIMIZE 的操作是将当前表复制到临时表操作后再删除当前表，最后将临时表改名

		
mysql> show processlist;
+--------+-----------------+---------------------+---------------------------+---------+-------+-----------------------------+--------------------------+
| Id     | User            | Host                | db                        | Command | Time  | State                       | Info                     |
+--------+-----------------+---------------------+---------------------------+---------+-------+-----------------------------+--------------------------+
|      1 | event_scheduler | localhost           | NULL                      | Daemon  |    14 | Waiting for next activation | NULL                     |
| 115835 | dba             | 192.168.6.20:49664  | example                   | Query   |     9 | copy to tmp table           | OPTIMIZE TABLE `mytable` |
| 115986 | dba             | localhost           | example                   | Query   |     0 | NULL                        | show processlist         |
+--------+-----------------+---------------------+---------------------------+---------+-------+-----------------------------+--------------------------+
17 rows in set (0.00 sec)
		
		

转换ENGINE从MyISAM到InnoDB会经历creating table然后copy to tmp table在修改表名几个阶段，过程非常缓慢

		
mysql> show processlist;
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
| Id   | User            | Host                | db      | Command | Time  | State                       | Info                                     |
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
|    1 | event_scheduler | localhost           | NULL    | Daemon  |    10 | Waiting for next activation | NULL                                     |
| 3167 | dba             | 192.168.6.20:56723  | example | Query   |     2 | creating table              | ALTER TABLE `mytable`	ENGINE=InnoDB   |
| 3172 | dba             | localhost           | example | Query   |     0 | NULL                        | show processlist                         |
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
18 rows in set (0.00 sec)
		
		

copy to tmp table 过程

		
mysql> show processlist;
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
| Id   | User            | Host                | db      | Command | Time  | State                       | Info                                     |
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
|    1 | event_scheduler | localhost           | NULL    | Daemon  |    21 | Waiting for next activation | NULL                                     |
| 3167 | dba             | 192.168.6.20:56723  | example | Query   |    13 | copy to tmp table           | ALTER TABLE `mytable`	ENGINE=InnoDB   |
| 3172 | dba             | localhost           | example | Query   |     0 | NULL                        | show processlist                         |
+------+-----------------+---------------------+---------+---------+-------+-----------------------------+------------------------------------------+
18 rows in set (0.00 sec)
		
		

此时我们查看mysql data目录会看到临时表文件

		
# ll /var/lib/mysql/hx9999_real_history/
		
-rw-rw---- 1 mysql mysql      9522 May 16 17:17 #sql-c2f_c5f.frm
-rw-rw---- 1 mysql mysql        48 May 16 17:17 #sql-c2f_c5f.par
-rw-rw---- 1 mysql mysql 637534208 May 16 17:29 #sql-c2f_c5f#P#p0.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p1.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p2.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p3.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p4.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p5.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p6.ibd
-rw-rw---- 1 mysql mysql    180224 May 16 17:17 #sql-c2f_c5f#P#p7.ibd		
		
		

使用各种手段保证select操作不被受阻，只要select一直可以查询网站前端就能提供80%的功能，一旦select受阻一切都是浮云。

保证 select 操作优先于其他操作

		
UPDATE [LOW_PRIORITY] [IGNORE] tbl_name  
SET col_name1=expr1 [, col_name2=expr2 ...]  
[WHERE where_definition]  
[ORDER BY ...]  
[LIMIT row_count]		
		
		

update的时候增加 LOW_PRIORITY 参数，可以降低更新语句的优先级。

my.cnf

		
[mysqld]		
low_priority_updates=1
		
		

或者启动是添加--low-priority-updates参数

全局开启

		
SET @@global.low_priority_updates = 1;		
		
		

适用于本次会话连接

		
SET @@session.low_priority_updates = 1;
		
		

使用 limit 限制更新记录的数量

		
update mytable set status='Y' where status='N' limit 1000;		
		
		

		
update mytable set status='Y',update_date=now(),op_user='neo' where status='N';