为了大家更容易理解我举出的SQL语句，本文假定已经建立了一个学生成绩管理数据库，全文均以学生成绩的管理为例来描述。
1.在查询结果中显示列名：
a.用as关键字：select name as '姓名' from students order by age
b.直接表示：select name '姓名' from students order by age
2.精确查找:
a.用in限定范围：select * from students where native in ('湖南', '四川')
b.between...and：select * from students where age between 20 and 30
c.“=”：select * from students where name = '李山'
d.like:select * from students where name like '李%' (注意查询条件中有“%”，则说明是部分匹配，而且还有先后信息在里面，即查找以“李”开头的匹配项。所以若查询有“李”的所有对象，应该命令：'%李%';若是第二个字为李，则应为'_李%'或'_李'或'_李_'。)
e.[]匹配检查符：select * from courses where cno like '[AC]%' (表示或的关系，与"in(...)"类似，而且"[]"可以表示范围，如：select * from courses where cno like '[A-C]%')
3.对于时间类型变量的处理
a.smalldatetime：直接按照字符串处理的方式进行处理，例如：
select * from students where birth > = '1980-1-1' and birth <= '1980-12-31'
select * from students where birth > = '1980-1-1' and birth <= '1980-12-31'
4.集函数
a.count()求和，如：select count(*) from students (求学生总人数)
b.avg(列)求平均，如：select avg(mark) from grades where cno=’B2’
c.max(列)和min(列)，求最大与最小
5.分组group
常用于统计时，如分组查总数：
select gender,count(sno)
from students
group by gender
(查看男女学生各有多少)
select gender,count(sno)
from students
group by gender
(查看男女学生各有多少)
注意：从哪种角度分组就从哪列"group by"
对于多重分组，只需将分组规则罗列。比如查询各届各专业的男女同学人数 ，那么分组规则有：届别(grade)、专业(mno)和性别(gender)，所以有"group by grade, mno, gender"
select grade, mno, gender, count(*)
from students
group by grade, mno, gender
from students
group by grade, mno, gender
通常group还和having联用，比如查询1门课以上不及格的学生，则按学号(sno)分类有：
select sno,count(*) from grades
where mark<60
group by sno
having count(*)>1
where mark<60
group by sno
having count(*)>1
6.UNION联合
合并查询结果，如：
SELECT * FROM students
WHERE name like ‘张%’
UNION [ALL]
SELECT * FROM students
WHERE name like ‘李%’
WHERE name like ‘张%’
UNION [ALL]
SELECT * FROM students
WHERE name like ‘李%’
7.多表查询
a.内连接
select g.sno,s.name,c.coursename
from grades g JOIN students s ON g.sno=s.sno
JOIN courses c ON g.cno=c.cno
(注意可以引用别名)
b.外连接
b1.左连接
select courses.cno,max(coursename),count(sno)
from courses LEFT JOIN grades ON courses.cno=grades.cno
group by courses.cno
from grades g JOIN students s ON g.sno=s.sno
JOIN courses c ON g.cno=c.cno
(注意可以引用别名)
b.外连接
b1.左连接
select courses.cno,max(coursename),count(sno)
from courses LEFT JOIN grades ON courses.cno=grades.cno
group by courses.cno
左连接特点：显示全部左边表中的所有项目，即使其中有些项中的数据未填写完全。
左外连接返回那些存在于左表而右表中却没有的行，再加上内连接的行。
b2.右连接
与左连接类似
b3.全连接
select sno,name,major
from students FULL JOIN majors ON students.mno=majors.mno
from students FULL JOIN majors ON students.mno=majors.mno
两边表中的内容全部显示
c.自身连接
select c1.cno,c1.coursename,c1.pno,c2.coursename
from courses c1,courses c2 where c1.pno=c2.cno
from courses c1,courses c2 where c1.pno=c2.cno
采用别名解决问题。
d.交叉连接
select lastname+firstname from lastname CROSS JOIN firstanme
相当于做笛卡儿积
8.嵌套查询
a.用关键字IN,如查询李山的同乡：
select * from students
where native in (select native from students where name=’ 李山’)
where native in (select native from students where name=’ 李山’)
b.使用关键字EXIST,比如，下面两句是等价的：
select * from students
where sno in (select sno from grades where cno=’B2’)
where sno in (select sno from grades where cno=’B2’)
select * from students where exists
(select * from grades where
grades.sno=students.sno AND cno=’B2’)
(select * from grades where
grades.sno=students.sno AND cno=’B2’)
9.关于排序order
a.对于排序order，有两种方法：asc升序和desc降序
b.对于排序order,可以按照查询条件中的某项排列，而且这项可用数字表示，如：
select sno,count(*) ,avg(mark) from grades
group by sno
having avg(mark)>85
order by 3
group by sno
having avg(mark)>85
order by 3
10.其他
a.对于有空格的识别名称，应该用"[]"括住。
b.对于某列中没有数据的特定查询可以用null判断，如select sno,courseno from grades where mark IS NULL
c.注意区分在嵌套查询中使用的any与all的区别，any相当于逻辑运算“||”而all则相当于逻辑运算“&&”
d.注意在做否定意义的查询是小心进入陷阱：
如，没有选修‘B2’课程的学生 ：
select students.*
from students, grades
where students.sno=grades.sno
AND grades.cno <> ’B2’
from students, grades
where students.sno=grades.sno
AND grades.cno <> ’B2’
上面的查询方式是错误的，正确方式见下方：
select * from students
where not exists (select * from grades
where grades.sno=students.sno AND cno='B2')
where not exists (select * from grades
where grades.sno=students.sno AND cno='B2')
11.关于有难度多重嵌套查询的解决思想：
如，选修了全部课程的学生：
select *
from students
where not exists ( select *
from courses
where NOT EXISTS
(select *
from grades
where sno=students.sno
AND cno=courses.cno))
from students
where not exists ( select *
from courses
where NOT EXISTS
(select *
from grades
where sno=students.sno
AND cno=courses.cno))
最外一重：从学生表中选，排除那些有课没选的。用not exist。由于讨论对象是课程，所以第二重查询从course表中找，排除那些选了课的即可。
本文转自loveme2351CTO博客，原文链接：http://blog.51cto.com/loveme23/8600
，如需转载请自行联系原作者
