码迷,mamicode.com
首页 > 数据库 > 详细

[oldboy-django][2深入django]mysql查询语句--原生sql

时间:2017-11-01 11:05:38      阅读:220      评论:0      收藏:0      [点我收藏+]

标签:复制   理解   color   upd   多列   max   each   一点   tar   

#1 增(一共有三种方式)

    # 插入单条记录
    insert into t1(name,...) values(lzp,..);
    注意一点:t1(name,...)必须包含所有非空列(除去自增列)

    # 插入多条记录
       insert into t1(name,age) values(lzp,12),(alex,30);

    # 拷贝复制(拷贝t2表的数据到t1)
       insert into t1(name,age) values select name,age from t2;

#2 改
    update t1 set name=b,age=13 where name=lzp;
    update t1 set name=c where name=alex;

#3 查(重点)

    select [...] from tb [...]

    从前面[]扩展:
        select name as cname, age from tb1; # 起别名
        select name, age, 1 from tb1;  # 多显示一列,该列的值为常量
        select name, age, variable from tb1; # 多显示一列,该列的值为变量,变量的定义在前面实现

    从后面的[]扩展:
    1 where
        select name, age from tb1 where id in2,4,6) # id =2 或id=4,或id=6 集合(2,4,6)
        select name, age from tb1 where id <> 3         # <>不等于(小于<或者大于>select name, age from tb1 where id not in (2,4,6)
        select name, age from tb1 where id between 5 and 12; # id in [5,12] #闭区间集合[5,12]
        select name, age from tb1 where id in (select id from tb2); #
        通配符%(多个任意字符), 下划线_(一个任意字符)
            select name, age from tb1 where name like "a%" # 以a开头的名字

        note: 不能允许这种 where id =1 and id=2, 这样查到的肯定是Null
    2 分页limit [start], count  ; start
        是从0开始,默认值=0;count表示查看的记录数
        select name, age from tb1 limit 10; #查看第一条数据到第10条数据;
        select name, age from tb1 limit 10,10; #查看第11条数据到第20条数据

        简单分页: page = int(input(‘‘).strip())
                  start = (page-1)*10
                  select name, age from tb1 limit start,10

         limit 10 offset=20; 等效

    3 order by 排序
        order by 列名 [asc|desc]  # 列名都是整数
        select name, age from tb1 order by id desc;# 按id从大到小排序

        select name, age from tb1 order by id desc limit 0,2;

        # 多列排序
        order by cow_1 desc, cow_2 desc

        # 注意不是order by cow_1, order by cow_2
        先按cow_1降序排,如果cow_1有相同的记录,对cow_1相同的记录按cow_2进行降序排

    4 分组group by + 聚合函数

        group by part_id: 可以理解成,将part_id相同的折成一条记录(怎么折?去重(最好不用distinct)
        对于相同的记录,可以根据函数count(),max(),min(),sum(),avg()折成一条记录
        这些函数称为聚合函数

        example1:
        用户表,部门表,查看每个部门有多少人
        select part_id, count(id), max(id) from tb1 group by part_id;

    5 分组group by + 聚合函数 + 二次刷选(having)
        对聚合函数结果进行二次刷选时,必须用having
        select part_id, count(id) from tb1 group by part_id having count(id) > 1
        # 刷选出部门人数大于1的部门


    6 连表操作
        将所有表select出来,然后告诉他们的关系
            select * from user, department where user.part_id = department.id

        推荐:left join ...on (之前版本性能更好,目前版本两者性能是一致)
            #将所有表select 出来; select * from tb1 left join tb2
            # on tb1.part_id = tb2.id 告诉他们表之间的关系

            select * from user left join department on user.part_id = department.id

    7 连表操作left join 和right join, inner join 区别
            select * from tb1 left join tb2 where ...
            tb1会全部显示,左边的表会全部显示;

            inner join 等效于将实现left join连表,后将Null出现的整行隐藏(不显示)


    8 多张表连接
            select * from tb1
                left join tb2 where ...  # 此时只能告诉tb1和tb2的关系
                left join tb3 where ...  # 此时可以告诉tb1和tb3的关系,tb2和tb3的关系(此时,已经将tb1,tb2,tb3关联进来了)

            连表的时候,如果遇到列名相同,*可变成表名.
                select tb1.id,
                       tb2.id,
                       tb3.id
                from tb1
                    left join tb2 where ..
                    left join tb3 where ..

            举例:  http://images2015.cnblogs.com/blog/425762/201608/425762-20160803224643778-2071849037.png
                    http://www.cnblogs.com/wupeiqi/articles/5729934.html
            学生,班级,老师,课程,成绩表:


            学生表:foreign key 班级表
            班级表:
            课程表:foreign key 老师
            成绩表:学生id, 课程id, 分数; foreign key 学生表, foreign key 课程表

    #9 临时表
        将查询得到的数据,作为一个临时表;在内存中存储,但并没有写到硬盘上。
        语法:
        (select * from tb1 where age > 20) as B;
        select sid fromselect * from tb1 where age > 20) as B;


    #10 增加显示列(前提:select score from tb1 where name="yuwen"只能是一个值)

        select id,  (select score from tb1 where name="yuwen") from tb2;
        tb1和tb2不是同一表,增加的显示列不是处理同一个表同一行的数据

    #11 增加显示列
        select id, (select score from tb1 as s2 where s2.id= s1.id) as score from
        tb1 as s1

        # 最外层的循环,一行一行数据
        # 里面的循环,每一行中的一列一列

    #12 条件语句
        case when min(num) < 10 then 0 else 1 end as c

    #13 mysql的三元运算
        if(True, 1,0)
        # 计算课程平均分从高到底显示,显示任课老师
        avg(if(isnull(score.num), 0, score.num))
        # 因为为空的时候无法计算

        select score.course_id, course.cname, teacher.tname, avg(if(isnull(score.num), 0, score.num)) as average from score
                left join course on score.course_id=course.cid
                left join teacher on course.teacher_id=teacher.tid
                group by score.course_id
                order by average desc;  # 此时可以用average,但是在select里面不能用average

    #14 双重循环
        select * from
            (select
                A.student_id,
                A.course_id,
                A.num,
                (select B.num from score as B where B.course_id = A.course_id order by B.num desc limit 0,1) as frist_s,
                (select B.num from score as B where B.course_id = A.course_id order by B.num desc limit 1,1) as second_s
            from score  as A)
        as C  where C.num >= C.second_s

    #15 多表操作,直接用连表,会比较好理解和设置where


    #16 # 查询没有学过李平老师课学生的姓名,id---没有学过李平老师任何一门课的学生
            先刷选出选过李平老师任意一门课的学生,

        # 碰到Not,一定要在最顶一层进行Not, 在学生表里面not in (学过老师课程学生的id,还有进行分组过滤掉重复的)
        # 先查到李平老师教过哪些课;

        -- select student.sid,student.sname from student where student.sid not in (
        --     select student_id from score
        --     where score.course_id in (select cid from course left join teacher on course.teacher_id=teacher.tid where teacher.tname=李平老师)
        --     group by student_id);


    #17  where  group by having : 先进行where一次刷选,再进行group having刷选


    #18  union 上下连表(union自动去重,union all则不会去重)
        select id, name from tb1
        union
        select sid, sname frou stb1;

 

[oldboy-django][2深入django]mysql查询语句--原生sql

标签:复制   理解   color   upd   多列   max   each   一点   tar   

原文地址:http://www.cnblogs.com/liuzhipenglove/p/7765392.html

(0)
(0)
   
举报
评论 一句话评论(0
登录后才能评论!
© 2014 mamicode.com 版权所有  联系我们:gaon5@hotmail.com
迷上了代码!