这里需要使用到case when语句，这个就是sql中的if else语句： select a.ins_id, b.product_id, b.product_name, c.cust_id, c.cust_name, c.cust_sex, c.cust_age, c.family_num, -- 男 -- 这个地方根据数据库字段的不同，处理方式也不同 -- 如果数据库中cust_sex的数据类型本身就是0和1，那么就不需要转换 -- 只列出来即可 (case when c.cust_sex=‘男‘ then 1 else 0 end) as f, -- 女 (case when c.cust_sex=‘女‘ then 1 else 0 end) as as m, -- 其他的依次类推 -- 家庭成员数 (case when c.family_num=1 then 1 else 0 end) as p_1, (case when c.family_num=2 then 1 else 0 end) as P_2, (case when c.family_num=3 then 1 else 0 end) as p_3, (case when c.family_num>3 then 1 else 0 end) as gt3, -- 客户年龄 (case when c.cust_age<=25 then 1 else 0 end) as lt25, (case when c.cust_age>25 and c.cust_age<=35 then 1 else 0 end) as gt25lt35, (case when c.cust_age>35 and c.cust_age<=45 then 1 else 0 end) as gt35lt45, (case when c.cust_age>45 and c.cust_age<=55 then 1 else 0 end) as gt45lt55, (case when c.cust_age>55 then 1 else 0 end) as gt55 from insurance a, product b, customer c where a.product_id=b.product_id and a.cust_id=c.cust_id 最后分组count一下即可： select a.product_id, a.product_name, count(a.ins_id) as ins_num, -- 性别 count(a.f) as f_num, count(a.m) as m_num, -- 成员数 count(a.p_1) as p_1_num, count(a.p_2) as p_1_num, count(a.p_3) as p_1_num, count(a.gt3) as gt3_num, -- 年龄 count(lt25) as lt25_num, count(gt25lt35) as gt25lt35_num, count(gt35lt45) as gt25lt35_num, count(gt45lt55) as gt25lt35_num, count(gt55) as gt55_num from( select a.ins_id, b.product_id, b.product_name, c.cust_id, c.cust_name, c.cust_sex, c.cust_age, c.family_num, -- 男 -- 这个地方根据数据库字段的不同，处理方式也不同 -- 如果数据库中cust_sex的数据类型本身就是0和1，那么就不需要转换 -- 只列出来即可 (case when c.cust_sex=‘男‘ then 1 else 0 end) as f, -- 女 (case when c.cust_sex=‘女‘ then 1 else 0 end) as as m, -- 其他的依次类推 -- 家庭成员数 (case when c.family_num=1 then 1 else 0 end) as p_1, (case when c.family_num=2 then 1 else 0 end) as P_2, (case when c.family_num=3 then 1 else 0 end) as p_3, (case when c.family_num>3 then 1 else 0 end) as gt3, -- 客户年龄 (case when c.cust_age<=25 then 1 else 0 end) as lt25, (case when c.cust_age>25 and c.cust_age<=35 then 1 else 0 end) as gt25lt35, (case when c.cust_age>35 and c.cust_age<=45 then 1 else 0 end) as gt35lt45, (case when c.cust_age>45 and c.cust_age<=55 then 1 else 0 end) as gt45lt55, (case when c.cust_age>55 then 1 else 0 end) as gt55 from insurance a, product b, customer c where a.product_id=b.product_id and a.cust_id=c.cust_id ) a group by b.product_id, b.product_name
到现在未知，我们所要的结果就完全出来了。整个过程是一个化繁为简，再由简单堆砌为复杂的过程。 对于初学者，培养出这样的思维方式似乎还很难，但是只要经过一两个这样的需求的练习，这中思维 方式就很容易形成了，到最后，当你看到一个报表结构的时候，这样的思维过程仅仅是一瞬间的事， 你的脑海里是n张的数据表格，经过相应关联之后，你的脑海里得到是报表结构的前一张结果集的结构， 然后再往前推前一张结果集的结构，直到推到主表，然后再正向推一编，最后推到完整的报表结构， 这个思维过程非常快，很可能再你的脑子只推一两步的时候，你就已经知道怎么写了。
谨以此文送给初学sql的朋友们。
原文地址：http://www.cnblogs.com/chengyihardworking/p/3810132.html
