Hive sql语句必练50题-入门到精通(2)(转载)

Hive sql语句必练50题-入门到精通(2)

 

原始链接：https://blog.csdn.net/Thomson617/article/details/83280617
原创Thomson617 最后发布于2018-10-22 23:08:54 阅读数 2546  收藏
展开
承接: Hive sql语句必练50题-入门到精通(1)

– 21、查询不同老师所教不同课程平均分从高到低显示:
– 方法1

select course.c_id,course.t_id,t_name,round(avg(s_score),2)as avgscore from course
    join teacher on teacher.t_id=course.t_id
    join score on course.c_id=score.c_id
    group by course.c_id,course.t_id,t_name order by avgscore desc;
1
2
3
4
– 方法2

select course.c_id,course.t_id,t_name,round(avg(s_score),2)as avgscore from course,teacher,score
   where teacher.t_id=course.t_id and course.c_id=score.c_id
    group by course.c_id,course.t_id,t_name order by avgscore desc;
1
2
3
– 22、查询所有课程的成绩第2名到第3名的学生信息及该课程成绩:

select tmp1.* from
    (select * from score where c_id='01' order by s_score desc limit 3)tmp1
    order by s_score asc limit 2
union all select tmp2.* from
    (select * from score where c_id='02' order by s_score desc limit 3)tmp2
    order by s_score asc limit 2
union all select tmp3.* from
    (select * from score where c_id='03' order by s_score desc limit 3)tmp3
    order by s_score asc limit 2;
1
2
3
4
5
6
7
8
9
– 23、统计各科成绩各分数段人数：课程编号,课程名称,[100-85],[85-70],[70-60],[0-60]及所占百分比

select c.c_id,c.c_name,tmp1.s0_60, tmp1.percentum,tmp2.s60_70, tmp2.percentum,tmp3.s70_85, tmp3.percentum,tmp4.s85_100, tmp4.percentum
from course c
join(select c_id,sum(case when s_score<60 then 1 else 0 end )as s0_60,
               round(100*sum(case when s_score<60 then 1 else 0 end )/count(c_id),2)as percentum
     from score group by c_id)tmp1 on tmp1.c_id =c.c_id
left join(select c_id,sum(case when s_score<70 and s_score>=60 then 1 else 0 end )as s60_70,
               round(100*sum(case when s_score<70 and s_score>=60 then 1 else 0 end )/count(c_id),2)as percentum
     from score group by c_id)tmp2 on tmp2.c_id =c.c_id
left join(select c_id,sum(case when s_score<85 and s_score>=70 then 1 else 0 end )as s70_85,
               round(100*sum(case when s_score<85 and s_score>=70 then 1 else 0 end )/count(c_id),2)as percentum
     from score group by c_id)tmp3 on tmp3.c_id =c.c_id
left join(select c_id,sum(case when s_score>=85 then 1 else 0 end )as s85_100,
               round(100*sum(case when s_score>=85 then 1 else 0 end )/count(c_id),2)as percentum
     from score group by c_id)tmp4 on tmp4.c_id =c.c_id;
1
2
3
4
5
6
7
8
9
10
11
12
13
14
– 24、查询学生平均成绩及其名次:

select tmp.*,row_number()over(order by tmp.avgScore desc) Ranking from
  (select student.s_id,
          student.s_name,
          round(avg(score.s_score),2) as avgScore
  from student join score
  on student.s_id=score.s_id
  group by student.s_id,student.s_name)tmp
order by avgScore desc;
1
2
3
4
5
6
7
8
– 25、查询各科成绩前三名的记录

–课程id为01的前三名

select score.c_id,course.c_name,student.s_name,s_score from score
join student on student.s_id=score.s_id
join course on  score.c_id='01' and course.c_id=score.c_id
order by s_score desc limit 3;  
1
2
3
4
–课程id为02的前三名

select score.c_id,course.c_name,student.s_name,s_score 
from score
join student on student.s_id=score.s_id
join course on  score.c_id='02' and course.c_id=score.c_id
order by s_score desc limit 3; 
1
2
3
4
5
–课程id为03的前三名

select score.c_id,course.c_name,student.s_name,s_score 
from score
join student on student.s_id=score.s_id
join course on  score.c_id='03' and course.c_id=score.c_id  
order by s_score desc limit 3;
1
2
3
4
5
– 26、查询每门课程被选修的学生数:

select c.c_id,c.c_name,tmp.number from course c
    join (select c_id,count(1) as number from score
        where score.s_score<60 group by score.c_id)tmp
    on tmp.c_id=c.c_id;
1
2
3
4
– 27、查询出只有两门课程的全部学生的学号和姓名:

select st.s_id,st.s_name from student st
  join (select s_id from score group by s_id having count(c_id) =2)tmp
    on st.s_id=tmp.s_id;
1
2
3
– 28、查询男生、女生人数:

select tmp1.man,tmp2.women from
    (select count(1) as man from student where s_sex='男')tmp1,
    (select count(1) as women from student where s_sex='女')tmp2;
1
2
3
– 29、查询名字中含有"风"字的学生信息:

select * from student where s_name like '%风%';
1
– 30、查询同名同性学生名单，并统计同名人数:

select s1.s_id,s1.s_name,s1.s_sex,count(*) as sameName
from student s1,student s2
where s1.s_name=s2.s_name and s1.s_id<>s2.s_id and s1.s_sex=s2.s_sex
group by s1.s_id,s1.s_name,s1.s_sex;
1
2
3
4
– 31、查询1990年出生的学生名单:

select * from student where s_birth like '1990%';
1
– 32、查询每门课程的平均成绩，结果按平均成绩降序排列，平均成绩相同时，按课程编号升序排列:

select score.c_id,c_name,round(avg(s_score),2) as avgScore from score
  join course on score.c_id=course.c_id
    group by score.c_id,c_name order by avgScore desc,score.c_id asc;
1
2
3
– 33、查询平均成绩大于等于85的所有学生的学号、姓名和平均成绩:

select score.s_id,s_name,round(avg(s_score),2)as avgScore from score
    join student on student.s_id=score.s_id
    group by score.s_id,s_name having avg(s_score) >= 85;
1
2
3
– 34、查询课程名称为"数学"，且分数低于60的学生姓名和分数:

select s_name,s_score as mathScore from student
    join (select s_id,s_score
            from score,course
            where score.c_id=course.c_id and c_name='数学')tmp
    on tmp.s_score < 60 and student.s_id=tmp.s_id;
1
2
3
4
5
– 35、查询所有学生的课程及分数情况:

select a.s_name,
    SUM(case c.c_name when '语文' then b.s_score else 0 end ) as chainese,
    SUM(case c.c_name when '数学' then b.s_score else 0 end ) as math,
    SUM(case c.c_name when '英语' then b.s_score else 0 end ) as english,
    SUM(b.s_score) as sumScore
  from student a
    join score b on a.s_id=b.s_id
    join course c on b.c_id=c.c_id
    group by s_name,a.s_id;
1
2
3
4
5
6
7
8
9
后续部分参见:
https://blog.csdn.net/Thomson617/article/details/83281254

Hive下的SQL经验总结：

(1).不支持非等值连接，一般使用left join、right join 或者inner join替代。
    •SQL中对两表内联可以写成：
        select * from dual a,dual b where a.key = b.key;
    •Hive中应为:
        select * from dual a join dual b on a.key = b.key; 
    而不是传统的格式：
        SELECT t1.a1 as c1, t2.b1 as c2 FROM t1, t2 WHERE t1.a2 = t2.b2    
        
(2).分号字符:不能智能识别concat(‘;’,key)，只会将‘；’当做SQL结束符号。
    •分号是SQL语句结束标记，在HiveQL中也是，但是在HiveQL中，对分号的识别没有那么智慧，例如：
        •select concat(key,concat(';',key)) from dual;
    •但HiveQL在解析语句时提示：
        FAILED: Parse Error: line 0:-1 mismatched input '<EOF>' expecting ) in function specification
    •解决的办法是，使用分号的八进制的ASCII码进行转义，那么上述语句应写成：
        •select concat(key,concat('\073',key)) from dual;

(3).不支持INSERT INTO 表 Values（）, UPDATE, DELETE等操作.这样的话，就不要很复杂的锁机制来读写数据。
    INSERT INTO syntax is only available starting in version 0.8。INSERT INTO就是在表或分区中追加数据。

(4).HiveQL中String类型的字段若是空(empty)字符串, 即长度为0, 那么对它进行IS NULL的判断结果是False，使用left join可以进行筛选行。

(5).不支持 ‘< dt <’这种格式的范围查找，可以用dt in(”,”)或者between替代。

(6).Hive不支持将数据插入现有的表或分区中，仅支持覆盖重写整个表，示例如下：
    INSERT OVERWRITE TABLE t1 SELECT * FROM t2;
    
(7).group by的字段,必须是select后面的字段，select后面的字段不能比group by的字段多.
    如果select后面有聚合函数,则该select语句中必须有group by语句;
    而且group by后面不能使用别名;
    有聚合函数存在就必须有group by.
    
(8).select , where 及 having 之后不能跟子查询语句(一般使用left join、right join 或者inner join替代)

(9).先join(及inner join) 然后left join或right join

(10).hive不支持group_concat方法,可用 concat_ws('|', collect_set(str)) 实现

(11).not in 和 <> 不起作用,可用left join tmp on tableName.id = tmp.id where tmp.id is null 替代实现

(12).hive 中‘不等于’不管是用！ 或者<>符号实现，都会将空值即null过滤掉，此时要用
        where （white_level<>'3' or  white_level is null） 
    或者 where (white_level!='3' or white_level is null )  来保留null 的情况。

(13).union all 后面的表不加括号,不然执行报错;
    hive也不支持顶层的union all，使用子查询来解决;
    union all 之前不能有DISTRIBUTE BY | SORT BY| ORDER BY | LIMIT 等查询条件