MySQL 下 ROW_NUMBER / DENSE_RANK / RANK 的实现
原文链接：http://hi.baidu.com/wangzhiqing999/item/7ca215d8ec9823ee785daa2b
MySQL 下 ROW_NUMBER / DENSE_RANK / RANK 的实现
CREATE TABLE test_rownum (
ID int,
Ke CHAR(1),
val INT
);
INSERT INTO test_rownum
SELECT 1, 'A', 1 UNION ALL
SELECT 2, 'A', 2 UNION ALL
SELECT 3, 'A', 3 UNION ALL
SELECT 4, 'B', 1 UNION ALL
SELECT 5, 'B', 2 UNION ALL
SELECT 6, 'B', 1 UNION ALL
SELECT 7, 'C', 3 UNION ALL
SELECT 8, 'C', 2 UNION ALL
SELECT 9, 'C', 2;
rownum 用于显示序号
ID 字段用于 标记 原有的序号位置.
普通的处理
等价于 ROW_NUMBER() OVER (ORDER BY ke, val)
SELECT
@rownum:=@rownum+1 AS rownum,
id,
ke,
val
FROM
(SELECT @rownum:=0) r,
test_rownum
ORDER BY
ke, val;
+--------+------+------+------+
| rownum | id | ke | val |
+--------+------+------+------+
| 1 | 1 | A | 1 |
| 2 | 2 | A | 2 |
| 3 | 3 | A | 3 |
| 4 | 4 | B | 1 |
| 5 | 6 | B | 1 |
| 6 | 5 | B | 2 |
| 7 | 8 | C | 2 |
| 8 | 9 | C | 2 |
| 9 | 7 | C | 3 |
+--------+------+------+------+
9 rows in set (0.00 sec)
PARTITION 的处理
等价于 ROW_NUMBER() OVER (PARTITION BY ke ORDER BY val)
SELECT
CASE
WHEN @ke != ke THEN @rownum:= 1
ELSE @rownum:= @rownum + 1
END AS rownum,
id,
@ke := ke AS ke,
val
FROM
(SELECT @ke:='') k,
(SELECT @rownum:=0) r,
test_rownum main
ORDER BY
ke, val;
+--------+------+------+------+
| rownum | id | ke | val |
+--------+------+------+------+
| 1 | 1 | A | 1 |
| 2 | 2 | A | 2 |
| 3 | 3 | A | 3 |
| 1 | 4 | B | 1 |
| 2 | 6 | B | 1 |
| 3 | 5 | B | 2 |
| 1 | 8 | C | 2 |
| 2 | 9 | C | 2 |
| 3 | 7 | C | 3 |
+--------+------+------+------+
9 rows in set (0.00 sec)
RANK DENSE_RANK 的处理
等价于 RANK() OVER (PARTITION BY ke ORDER BY val)
等价于 DENSE_RANK() OVER (PARTITION BY ke ORDER BY val)
SELECT
CASE
WHEN @ke != ke THEN @rownum:= 1
ELSE @rownum:= @rownum + 1
END AS ROW_NUMBER,
CASE
WHEN @ke != ke THEN @rank:= 1
WHEN @val = val THEN @rank
ELSE @rank:= @rownum
END AS RANK,
CASE
WHEN @ke != ke THEN @dense_rank:= 1
WHEN @val = val THEN @dense_rank
ELSE @dense_rank:= @dense_rank + 1
END AS DENSE_RANK,
id,
@ke := ke AS ke,
@val := val AS val
FROM
(SELECT @ke:='') k,
(SELECT @val:=0) v,
(SELECT @rownum:=0) r,
(SELECT @rank:=0) r2,
(SELECT @dense_rank:=0) d,
test_rownum main
ORDER BY
ke, val;
+------------+------+------------+------+------+------+
| ROW_NUMBER | RANK | DENSE_RANK | id | ke | val |
+------------+------+------------+------+------+------+
| 1 | 1 | 1 | 1 | A | 1 |
| 2 | 2 | 2 | 2 | A | 2 |
| 3 | 3 | 3 | 3 | A | 3 |
| 1 | 1 | 1 | 4 | B | 1 |
| 2 | 1 | 1 | 6 | B | 1 |
| 3 | 3 | 2 | 5 | B | 2 |
| 1 | 1 | 1 | 8 | C | 2 |
| 2 | 1 | 1 | 9 | C | 2 |
| 3 | 3 | 2 | 7 | C | 3 |
+------------+------+------------+------+------+------+
9 rows in set (0.00 sec)
=======================================================
两表关联情况下的测试.
CREATE TABLE test_rownum_main (
ID int,
Ke CHAR(1)
);
CREATE TABLE test_rownum_sub (
KeID int,
val INT
);
INSERT INTO test_rownum_main
SELECT 1, 'A' UNION ALL
SELECT 2, 'B' UNION ALL
SELECT 3, 'C';
INSERT INTO test_rownum_sub
SELECT 1, 1 UNION ALL
SELECT 1, 2 UNION ALL
SELECT 1, 3 UNION ALL
SELECT 2, 1 UNION ALL
SELECT 2, 2 UNION ALL
SELECT 2, 1 UNION ALL
SELECT 3, 3 UNION ALL
SELECT 3, 2 UNION ALL
SELECT 3, 2;
普通的处理
等价于 ROW_NUMBER() OVER (ORDER BY ke, val)
SELECT
@rownum:=@rownum+1 AS rownum,
main.id,
main.ke,
sub.val
FROM
test_rownum_main main
JOIN test_rownum_sub sub
ON (main.ID = sub.KeID),
(SELECT @rownum:=0) r
ORDER BY
ke, val;
+--------+------+------+------+
| rownum | id | ke | val |
+--------+------+------+------+
| 1 | 1 | A | 1 |
| 2 | 1 | A | 2 |
| 3 | 1 | A | 3 |
| 4 | 2 | B | 1 |
| 6 | 2 | B | 1 |
| 5 | 2 | B | 2 |
| 8 | 3 | C | 2 |
| 9 | 3 | C | 2 |
| 7 | 3 | C | 3 |
+--------+------+------+------+
9 rows in set (0.00 sec)
rownum 顺序不满足预期结果了。
可能是因为执行关联的时候， 就把 @rownum 计算了。
修改SQL语句.
SELECT
@rownum:=@rownum+1 AS rownum,
id,
ke,
val
FROM
(SELECT @rownum:=0) r,
(
SELECT
main.id,
main.ke,
sub.val
FROM
test_rownum_main main
JOIN test_rownum_sub sub
ON (main.ID = sub.KeID)
) subQuery
ORDER BY
ke, val;
+--------+------+------+------+
| rownum | id | ke | val |
+--------+------+------+------+
| 1 | 1 | A | 1 |
| 2 | 1 | A | 2 |
| 3 | 1 | A | 3 |
| 4 | 2 | B | 1 |
| 5 | 2 | B | 1 |
| 6 | 2 | B | 2 |
| 7 | 3 | C | 2 |
| 8 | 3 | C | 2 |
| 9 | 3 | C | 3 |
+--------+------+------+------+
9 rows in set (0.00 sec)
PARTITION 的处理
等价于 ROW_NUMBER() OVER (PARTITION BY ke ORDER BY val)
SELECT
CASE
WHEN @ke != ke THEN @rownum:= 1
ELSE @rownum:= @rownum + 1
END AS rownum,
id,
@ke := ke AS ke,
val
FROM
(SELECT @ke:='') k,
(SELECT @rownum:=0) r,
(
SELECT
main.id,
main.ke,
sub.val
FROM
test_rownum_main main
JOIN test_rownum_sub sub
ON (main.ID = sub.KeID)
) subQuery
ORDER BY
ke, val;
+--------+------+------+------+
| rownum | id | ke | val |
+--------+------+------+------+
| 1 | 1 | A | 1 |
| 2 | 1 | A | 2 |
| 3 | 1 | A | 3 |
| 1 | 2 | B | 1 |
| 2 | 2 | B | 1 |
| 3 | 2 | B | 2 |
| 1 | 3 | C | 2 |
| 2 | 3 | C | 2 |
| 3 | 3 | C | 3 |
+--------+------+------+------+
9 rows in set (0.00 sec)
DENSE_RANK 的处理
RANK 的处理
等价于 DENSE_RANK() OVER (PARTITION BY ke ORDER BY val)
等价于 RANK() OVER (PARTITION BY ke ORDER BY val)
SELECT
CASE
WHEN @ke != ke THEN @rownum:= 1
ELSE @rownum:= @rownum + 1
END AS ROW_NUMBER,
CASE
WHEN @ke != ke THEN @rank:= 1
WHEN @val = val THEN @rank
ELSE @rank:= @rownum
END AS RANK,
CASE
WHEN @ke != ke THEN @dense_rank:= 1
WHEN @val = val THEN @dense_rank
ELSE @dense_rank:= @dense_rank + 1
END AS DENSE_RANK,
id,
@ke := ke AS ke,
@val := val AS val
FROM
(SELECT @ke:='') k,
(SELECT @val:=0) v,
(SELECT @rownum:=0) r,
(SELECT @rank:=0) r2,
(SELECT @dense_rank:=0) d,
(
SELECT
main.id,
main.ke,
sub.val
FROM
test_rownum_main main
JOIN test_rownum_sub sub
ON (main.ID = sub.KeID)
) subQuery
ORDER BY
ke, val;
+------------+------+------------+------+------+------+
| ROW_NUMBER | RANK | DENSE_RANK | id | ke | val |
+------------+------+------------+------+------+------+
| 1 | 1 | 1 | 1 | A | 1 |
| 2 | 2 | 2 | 1 | A | 2 |
| 3 | 3 | 3 | 1 | A | 3 |
| 1 | 1 | 1 | 2 | B | 1 |
| 2 | 1 | 1 | 2 | B | 1 |
| 3 | 3 | 2 | 2 | B | 2 |
| 1 | 1 | 1 | 3 | C | 2 |
| 2 | 1 | 1 | 3 | C | 2 |
| 3 | 3 | 2 | 3 | C | 3 |
+------------+------+------------+------+------+------+
9 rows in set (0.00 sec)
MySQL 下 ROW_NUMBER / DENSE_RANK / RANK 的实现的更多相关文章
- SQL2005四个排名函数（row_number、rank、dense_rank和ntile）的比较
排名函数是SQL Server2005新加的功能.在SQL Server2005中有如下四个排名函数: .row_number .rank .dense_rank .ntile 下面分别介绍一下这四个 ...
- ROW_NUMBER、RANK、DENSE_RANK的用法
--NND,索性把2005的一起帖出来. ROW_NUMBER.RANK.DENSE_RANK的用法 (爱新觉罗.毓华 -- 广东深圳) SQL Server 引入几个新的排序(排名)函数,如ROW_ ...
- 四个排名函数（row_number、rank、dense_rank和ntile）的比较
排名函数是SQL Server2005新加的功能.在SQL Server2005中有如下四个排名函数: 1.row_number 2.rank 3.dense_rank 4.ntile 下面分别介绍一 ...
- SQL中的排名函数（ROW_NUMBER、RANK、DENSE_RANK、NTILE）简介
排名函数是Sql Server2005新增的功能,下面简单介绍一下他们各自的用法和区别. 在使用排名函数的时候需要注意以下三点: 1.排名函数必须有 OVER 子句. 2.排名函数必须有包含 ORDE ...
- Sql 四大排名函数（ROW_NUMBER、RANK、DENSE_RANK、NTILE）（转载）
Sql 四大排名函数(ROW_NUMBER.RANK.DENSE_RANK.NTILE)简介 排名函数是Sql Server2005新增的功能,下面简单介绍一下他们各自的用法和区别.我们新建一张O ...
- hive笔记之row_number、rank、dense_rank
hive中有三个与分组排序相关的分析函数(我起初也认为是窗口函数,后来看到手册里是把他们划到了Analytics functions下),row_number.rank.dense_rank,我一直傻 ...
- SQL With As 用法Sql 四大排名函数（ROW_NUMBER、RANK、DENSE_RANK、NTILE）简介
Sql 四大排名函数(ROW_NUMBER.RANK.DENSE_RANK.NTILE)简介 排名函数是Sql Server2005新增的功能,下面简单介绍一下他们各自的用法和区别.我们新建一张O ...
- sql server 2000 对应 sql server 2005的row_number()、rank()、DENSE_RANK( )、ntile( )等用法
转自CSDN:http://blog.csdn.net/htl258/article/details/4006717 SQL server 2005新增的几个函数,分别是row_number( ).r ...
- row_number()、rank()、dense_rank()、ntile()
原文:row_number().rank().dense_rank().ntile() SQL2005中row_number()等函数的用法 2005比2000新增了几个函数,分别是row_numbe ...
随机推荐
- web服务器、wsgi、应用
1.web服务器(比如Apache.Nginx.Tomcat.Jetty.IIS)属于静态服务器,只能返回静态的文件,服务器多是C/C++语言开发的.web服务器可以说一个request proxy, ...
- depth linear
float ConvertDepth( float depthFromTex, float4 cameraParams ){ const float near = cameraParams.z; co ...
- HTML5的测试总结
HTML5其实也是web的一种,所以基本的web测试的一些重点,HTML5上都要过一遍,不过也有其特殊之处. [需求设计测试] 需求是否合理.是否有更好的实现方法或者功能的遗漏,以及原型图测试,包括用 ...
- ElasticSearch 分布式集群
1.前言 Elasticsearch用于构建高可用和可扩展的系统.扩展的方式可以是购买更好的服务器(纵向扩展(vertical scale or scaling up))或者购买更多的服务器(横向扩展 ...
- 在ecshop中添加页面，并且实现后台管理
后台一共需要修改下面的四个文件 admin/template.php admin/includes/lib_template.php languages/zh_cn/admin/template.ph ...
- OpenSSL Heartbleed "心脏滴血"漏洞简单攻击示例
转自:http://www.lijiejie.com/openssl-heartbleed-attack/ OpenSSL Heartbleed漏洞的公开和流行让许多人兴奋了一把,也让另一些人惊慌了一 ...
- ionic - 运行起来
更新时间: 2018-8-1 (首次更新) 1.首先下载python(至于为什么安装,看截图) https://www.python.org/downloads/release/python-370/ ...
- textarea限制字符输入方法
function check(obj){ var Maxchar=20; if(obj.value.length>Maxchar){ //如果超出 obj.value=obj.value.sub ...
- MySQL 5.6.26几种安装包的差别
http://downloads.mysql.com/archives/community/ 一.MySQL Installer 5.6.26 mysql-installer-community-5. ...
- Oracle 字符串不为空条件
Oracle 中,空字符串存入到Oracle中会自动转换为NULL,另外VARCHAR2把空串等同于null处理. SQL from dual where null=null; 没有查到记录 SQL ...