Rewriting MySQL query without using GROUP BY - mysql

Question

Here is the table information:
Table name is Teaches,
+-----------+--------------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+-----------+--------------+------+-----+---------+-------+
| ID | varchar(5) | NO | PRI | NULL | |
| course_id | varchar(8) | NO | PRI | NULL | |
| sec_id | varchar(8) | NO | PRI | NULL | |
| semester | varchar(6) | NO | PRI | NULL | |
| year | decimal(4,0) | NO | PRI | NULL | |
+-----------+--------------+------+-----+---------+-------+
The requirement is to find which course appeared more than once in 2009(ID is the id of teachers)
Here is my query using GROUP BY:
select course_id
from teaches
where year= 2009
group by course_id
having count(id) >= 2;
How could I write this without using GROUP BY?

Answer 1

You may try this:
SELECT DISTINCT
T.course_id
FROM
teaches T
WHERE
T.course_id NOT IN (
SELECT
T1.course_id
FROM teaches AS T1 INNER JOIN teaches AS T2 ON T1.course_id = T2.course_id
AND T1.`year` = T2.`year`
AND T1.id <> T2.id
WHERE T1.`year` = 2009
);
Test Schema And Data:
DROP TABLE IF EXISTS `teaches`;
CREATE TABLE `teaches` (
`ID` varchar(5) CHARACTER SET utf8 DEFAULT NULL,
`course_id` varchar(8) CHARACTER SET utf8 DEFAULT NULL,
`sec_id` varchar(8) CHARACTER SET utf8 DEFAULT NULL,
`semester` varchar(6) CHARACTER SET utf8 DEFAULT NULL,
`year` decimal(4,0) DEFAULT NULL
);
INSERT INTO `teaches` VALUES ('66', '100', 'B', '11', '2009');
INSERT INTO `teaches` VALUES ('71', '100', 'A', '11', '2009');
INSERT INTO `teaches` VALUES ('64', '102', 'C', '12', '2010');
INSERT INTO `teaches` VALUES ('77', '102', 'B', '22', '2009');
Expected Output:
course_id
102
SQL FIDDLE DEMO

Answer 2

You could use something like this
select t.course_id
from (select t.*, count(*) as cnt from teaches t where year= 2009 ) t
where cnt > 1
Try to use stack overflow to search before posting a question,
Select Rows that appear more than once

Answer 3

Something like below should work:
select distinct course_id
from teaches o
where (select count(1) from teaches i where i.course_id = o.course_id and i.[year] = 2009) > 1

Answer 4

For your homework, below sql can be done.
This is followed your logic, id exist more than once is means course appeared more than once.
select DISTINCT T1.course_id
from teaches T1
where T1.course_id not in (
select a.course_id
from teaches as a inner join teaches as b
on a.course_id = b.course_id and a.year = b.year and a.id <> b.id
where a.year= 2009 )