要得到一组数据的中位数(例如某个地区或某家公司的收入中位数),我们一般要将这一任务细分为 3 个小任务:
1.将数据排序,并给每一行数据给出其在所有数据中的排名;
2.找出中位数的排名数字;
3.找出中间排名对应的值;
下面以某公司员工月收入为例,示例 MySQL 的一些复杂语句的使用。
方法一
创建测试表
首先创建一个收入表,建表语句为:
CREATE TABLE IF NOT EXISTS `employee` ( `id` INT AUTO_INCREMENT PRIMARY KEY, `name` VARCHAR(10) NOT NULL DEFAULT '', `income` INT NOT NULL DEFAULT '0' ) ENGINE = InnoDB DEFAULT CHARSET = utf8; INSERT INTO `employee` (`name`, `income`) VALUES ('麻子', 20000); INSERT INTO `employee` (`name`, `income`) VALUES ('李四', 12000); INSERT INTO `employee` (`name`, `income`) VALUES ('张三', 10000); INSERT INTO `employee` (`name`, `income`) VALUES ('王二', 16000); INSERT INTO `employee` (`name`, `income`) VALUES ('土豪', 40000);
完成任务 1
将数据排序,并给每一行数据给出其在所有数据中的排名:
SELECT t1.name, t1.income, COUNT(*) AS rank FROM employee AS t1, employee AS t2 WHERE t1.income < t2.income OR (t1.income = t2.income AND t1.name <= t2.name) GROUP BY t1.name, t1.income ORDER BY rank;
查询结果为:
完成小任务 2
找出中位数的排名数字:
SELECT (COUNT(*) + 1) DIV 2 as rank FROM employee;
查询结果为:
完成小任务 3
SELECT income AS median FROM (SELECT t1.name, t1.income, COUNT(*) AS rank FROM employee AS t1, employee AS t2 WHERE t1.income < t2.income OR (t1.income = t2.income AND t1.name <= t2.name) GROUP BY t1.name, t1.income ORDER BY rank) t3 WHERE rank = (SELECT (COUNT(*) + 1) DIV 2 FROM employee)
查询结果为:
至此,我们就找到了如何从一组数据中获得中位数的方法。
方法二
下面,来介绍另外一种优化排名语句的方法。
我们都知道如何给一组数据做排序操作,在本例中,实现方法如下:
SELECT name, income FROM employee ORDER BY income DESC
查询结果为:
那我们可不可以更进一步,对查询出的结果加一列,这一列的数据为排名呢?
我们可以通过 3 个自定义变量的方法来实现这一目标:
第一个变量用来记录当前行数据的收入
第二个变量用来记录上一行数据的收入
第三个变量用来记录当前行数据的排名
SET @curr_income := 0; SET @prev_income := 0; SET @rank := 0; SELECT `name`, @curr_income := income AS income, @rank := if(@prev_income != @curr_income, @rank + 1, @rank) AS rank, @prev_income := @curr_income AS dummy FROM employee ORDER BY income DESC
查询结果如下:
然后再找出中位数的排名数字,进一步找出收入的中位数:
SET @curr_income := 0; SET @prev_income := 0; SET @rank := 0; SELECT income AS median FROM (SELECT `name`, @curr_income := income AS income, @rank := if(@prev_income != @curr_income, @rank + 1, @rank) AS rank, @prev_income := @curr_income AS dummy FROM employee ORDER BY income DESC) AS t1 WHERE t1.rank = (SELECT (COUNT(*) + 1) DIV 2 FROM employee)
查询结果为:
至此,我们找了两种方法来解决中位数的问题。撒花。
推荐:《mysql教程》
以上就是在 MySQL 中,如何计算一组数据的中位数的详细内容,更多请关注名铺123其它相关文章!