SQL65 异常的邮件概率

 

条件

· 有一个邮件(email)表，id为主键， type是枚举类型，枚举成员为(completed，no_completed)，completed代表邮件发送是成功的，no_completed代表邮件是发送失败的

· 用户(user)表，id为主键，is_blacklist为0代表为正常用户，is_blacklist为1代表为黑名单用户

 

要求

1、每一个日期里面，正常用户发送给正常用户邮件失败的概率

2、结果保留到小数点后面3位(3位之后的四舍五入)

3、按照日期升序排序

上面例子查询结果如下:

 

解析

根据题目要求，需要求正常用户发送到正常用户的失败率，所以先连接user表求出正常用户

select xxx
from email as e 
join user as u1 on e.send_id = u1.id and u1.is_blacklist = 0
join user as u2 on e.receive_id = u2.id and u2.is_blacklist = 0

然后求出失败的概率 = 失败的次数 / 总次数

总次数为count(e.type)；失败的次数可以使用case..when..函数去计算，得sum (case when e.type = completed then 0 else 1 end)

select round(1.0 * sum(case when e.type = 'completed' then 0 else 1 end) / count(e.type), 3) as p
from email as e 
join user as u1 on e.send_id = u1.id and u1.is_blacklist = 0
join user as u2 on e.receive_id = u2.id and u2.is_blacklist = 0

联立其它条件可得

select e.date, round(1.0 * sum(case when e.type = 'completed' then 0 else 1 end) / count(e.type), 3) as p
from email as e 
join user as u1 on e.send_id = u1.id and u1.is_blacklist = 0
join user as u2 on e.receive_id = u2.id and u2.is_blacklist = 0
group by e.date
order by e.date;

 

标签：概率,u1,SQL65,u2,blacklist,user,type,id,邮件
来源： https://www.cnblogs.com/Panzm123/p/15254356.html