hive sql練習

參考

-- 創建學生表,學號,姓名,性別,年齡,班級
DROP TABLE student;
CREATE TEMPORARY TABLE  student(
  Sno int,
  Sname VARCHAR(20),
  Sex VARCHAR(20),
  Sage int,
  Sdept VARCHAR(20)
);

-- 創建 課程表, 課程編號,課程名稱
DROP TABLE course;
create TEMPORARY TABLE course(
  Cno int,
  Cname VARCHAR(20)
);
-- 創建 學生 課程關聯表,學號,課號,年級
create table sc(
  Sno int,
  Cno int,
  Grade int
);

-- 插入課程數據
INSERT into course values(1, 'database');
INSERT into course values(2, 'math');
INSERT into course values(3, 'info system');
INSERT into course values(4, 'system');
INSERT into course values(5, 'data structure');
INSERT into course values(6, 'data deal');

-- 插入學生選課數據和成績
INSERT into sc values(95001,1,81);
INSERT into sc values(95001,2,85);
INSERT into sc values(95001,3,88);
INSERT into sc values(95001,4,70);
INSERT into sc values(95002,2,90);
INSERT into sc values(95002,3,80);
INSERT into sc values(95002,4,71);
INSERT into sc values(95002,5,60);
INSERT into sc values(95003,1,82);
INSERT into sc values(95003,3,90);
INSERT into sc values(95003,5,100);
INSERT into sc values(95004,1,80);
INSERT into sc values(95004,2,92);
INSERT into sc values(95004,4,91);
INSERT into sc values(95004,5,70);
INSERT into sc values(95005,1,70);
INSERT into sc values(95005,2,92);
INSERT into sc values(95005,3,99);
INSERT into sc values(95005,6,87);
INSERT into sc values(95006,1,72);
INSERT into sc values(95006,2,62);
INSERT into sc values(95006,3,100);
INSERT into sc values(95006,4,59);
INSERT into sc values(95006,5,60);
INSERT into sc values(95006,6,98);
INSERT into sc values(95007,3,68);
INSERT into sc values(95007,4,91);
INSERT into sc values(95007,5,94);
INSERT into sc values(95007,6,78);
INSERT into sc values(95008,1,98);
INSERT into sc values(95008,3,89);
INSERT into sc values(95008,6,91);
INSERT into sc values(95009,2,81);
INSERT into sc values(95009,4,89);
INSERT into sc values(95009,6,100);
INSERT into sc values(95010,2,98);
INSERT into sc values(95010,5,90);
INSERT into sc values(95010,6,80);
INSERT into sc values(95011,1,81);
INSERT into sc values(95011,2,91);
INSERT into sc values(95011,3,81);
INSERT into sc values(95011,4,86);
INSERT into sc values(95012,1,81);
INSERT into sc values(95012,3,78);
INSERT into sc values(95012,4,85);
INSERT into sc values(95012,6,98);
INSERT into sc values(95013,1,98);
INSERT into sc values(95013,2,58);
INSERT into sc values(95013,4,88);
INSERT into sc values(95013,5,93);
INSERT into sc values(95014,1,91);
INSERT into sc values(95014,2,100);
INSERT into sc values(95014,4,98);
INSERT into sc values(95015,1,91);
INSERT into sc values(95015,3,59);
INSERT into sc values(95015,4,100);
INSERT into sc values(95015,6,95);
INSERT into sc values(95016,1,92);
INSERT into sc values(95016,2,99);
INSERT into sc values(95016,4,82);
INSERT into sc values(95017,4,82);
INSERT into sc values(95017,5,100);
INSERT into sc values(95017,6,58);
INSERT into sc values(95018,1,95);
INSERT into sc values(95018,2,100);
INSERT into sc values(95018,3,67);
INSERT into sc values(95018,4,78);
INSERT into sc values(95019,1,77);
INSERT into sc values(95019,2,90);
INSERT into sc values(95019,3,91);
INSERT into sc values(95019,4,67);
INSERT into sc values(95019,5,87);
INSERT into sc values(95020,1,66);
INSERT into sc values(95020,2,99);
INSERT into sc values(95020,5,93);
INSERT into sc values(95021,2,93);
INSERT into sc values(95021,5,91);
INSERT into sc values(95021,6,99);
INSERT into sc values(95022,3,69);
INSERT into sc values(95022,4,93);
INSERT into sc values(95022,5,82);
INSERT into sc values(95022,6,100);

INSERT into student values(95001,"liyong","M",20,"CS");
INSERT into student values(95002,"licheng","F",19,"IS");
INSERT into student values(95003,"wangming","F",22,"MA");
INSERT into student values(95004,"zhangli","M",19,"IS");
INSERT into student values(95005,"liugang","M",18,"MA");
INSERT into student values(95006,"sunqing","M",23,"CS");
INSERT into student values(95007,"yisiling","F",19,"MA");
INSERT into student values(95008,"lina","F",18,"CS");
INSERT into student values(95009,"mengyuanyuan","F",18,"MA");
INSERT into student values(95010,"kongxiaotao","M",19,"CS");
INSERT into student values(95011,"baoxiaobo","M",18,"MA");
INSERT into student values(95012,"sunhua","F",20,"CS");
INSERT into student values(95013,"fengwei","M",21,"CS");
INSERT into student values(95014,"wangxiaoli","F",19,"CS");
INSERT into student values(95015,"wangjun","M",18,"MA");
INSERT into student values(95016,"qianguo","M",21,"MA");
INSERT into student values(95017,"wangfengjuan","F",18,"IS");
INSERT into student values(95018,"wangyi","F",19,"IS");
INSERT into student values(95019,"xingxiaoli","F",19,"IS");
INSERT into student values(95020,"zhaoqian","M",21,"IS");
INSERT into student values(95021,"zhouer","M",17,"MA");
INSERT into student values(95022,"zhangming","M",20,"MA");

SELECT * FROM course;

-- 查詢學生總人數
SELECT count(*) FROM student;

-- 查詢全體學生的學號與姓名
SELECT student.Sno, student.Sname FROM student;

-- 查詢選修了課程的學生姓名,
SELECT student.Sname FROM student
JOIN sc
on student.Sno = sc.Sno
GROUP BY student.Sname;

-- 計算1號課程的學生平均成績,sc表
SELECT avg(sc.Grade) FROM sc WHERE Cno = 1 GROUP BY Cno;

-- 這一個寫法要快一點
select avg(Grade) from sc group by Cno having Cno=1;

-- 查詢各科成績平均分
-- 這個所有學系的各科平均分
SELECT course.Cname, avg(sc.Grade) FROM course
JOIN sc
on course.Cno=sc.Cno
GROUP BY course.Cname;

-- 這樣寫要快一點
SELECT course.Cname, avg(sc.Grade) FROM course
JOIN sc
on course.Cno=sc.Cno
GROUP BY course.Cname,sc.Cno;

-- 查詢選修1號課程的學生最高分數
SELECT max(sc.Grade) FROM sc WHERE sc.Cno = 1;


-- 這個寫法要快一點
select Grade from sc where Cno=1 order by Grade desc limit 1;

select Grade from sc where Cno=1 distribute by Grade sort by Grade desc limit 1;

-- 求各個課程號及相應的選課人數
SELECT sc.Cno, count(1) as num from sc GROUP BY Cno;

-- 查詢選修了3門以上的課程的學生學號
SELECT sc.Sno FROM sc GROUP BY Sno HAVING count(1) > 3;

-- 查詢學生信息,結果按學號全局有序
SELECT student.* FROM student ORDER BY student.Sno;


-- 查詢學生信息,結果區分性別按年齡有序
SELECT * FROM student ORDER BY sex,Sage asc;

-- 查詢每個學生及其選修課程的成績情況
SELECT student.*,course.Cname,sc.Grade FROM student
JOIN sc
on sc.Sno = student.Sno
JOIN course
on course.Cno=sc.Cno
ORDER BY student.Sno;

-- 查詢學生的得分情況
SELECT student.Sname,course.Cname,sc.Grade FROM student
JOIN sc
on sc.Sno = student.Sno
JOIN course
on course.Cno=sc.Cno
ORDER BY student.Sname;

-- 查詢選修2號課程且成績在90分以上的所有學生。
SELECT student.* FROM student
JOIN sc
on student.Sno=sc.Sno
WHERE sc.Cno=2 and sc.Grade > 90;

-- 查詢所有學生的信息,如果在成績表中有成績,則輸出成績表中的課程號
SELECT student.*,sc.Cno FROM student
left JOIN sc
on student.Sno=sc.Sno;

-- 查詢與“licheng”在同一個系學習的學生
-- exists寫法於join差不多快
SELECT stu1.Sname FROM student stu1 WHERE exists
(
  SELECT * FROM student stu2 WHERE stu2.Sname = 'licheng' and stu1.Sdept = stu2.Sdept
);

-- in 寫法
SELECT stu1.Sname FROM student stu1 WHERE stu1.Sdept in
(
  SELECT stu2.Sdept FROM student stu2 WHERE stu2.Sname = 'licheng'
);

-- join寫法最快
SELECT stu.Sname FROM student stu
JOIN student st
on st.Sname = 'licheng' and st.Sdept = stu.Sdept and stu.Sname != 'licheng';

select s2.* from student as s1 inner join student as s2 on(s2.Sdept=s1.Sdept) where s1.Sname='licheng';

原先數據格式是這樣,寫了個腳本轉成insert 形式

ss = '''95001,李勇,男,20,CS
95002,劉晨,女,19,IS
95003,王敏,女,22,MA
95004,張立,男,19,IS
95005,劉剛,男,18,MA
95006,孫慶,男,23,CS
95007,易思玲,女,19,MA
95008,李娜,女,18,CS
95009,夢圓圓,女,18,MA
95010,孔小濤,男,19,CS
95011,包小柏,男,18,MA
95012,孫花,女,20,CS
95013,馮偉,男,21,CS
95014,王小麗,女,19,CS
95015,王君,男,18,MA
95016,錢國,男,21,MA
95017,王風娟,女,18,IS
95018,王一,女,19,IS
95019,邢小麗,女,19,IS
95020,趙錢,男,21,IS
95021,週二,男,17,MA
95022,鄭明,男,20,MA
'''
rr = "INSERT into student values(%s);";
for el in ss.split('\n'):
    temp = ""
    i = 0
    for ell in el.split(','):
        if i == 0 or i==3:
            temp +=ell+','
        elif i == 1 or i == 2:
            temp +="\"%s\""%ell+','
        else:
            temp += "\"%s\""%ell
        i += 1
        
    print (rr%temp)
發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章