mysql count on multiple columns from another table - mysql

Question

I would like to get count of two columns from another table. For example, the source data is as below for the table called employee_master
emp_code dept_id sub_dept_id
1 10 22
2 11 20
3 10 22
4 10 22
5 11 20
6 10 21
Another source table called punching_data_table as shown below which is being joined with employee_master.
emp_code present
1 10:01 AM
2 10:02 AM
3 10:02 AM
4 10:03 AM
5 10:11 AM
6 10:09 AM //biometric punching time shown
The result is suppose to look like below.
dept_id dept_count sub_dept_id sub_dept_count
10 4 22 3
10 4 21 1
11 2 20 2
Tried with following code but it doesn't count seperately for both columns even though applied group by employee_master.dept_id, employee_master.sub_dept_id. Please note that there is only 1 column common between both table that is emp_code.
SELECT p.emp_code
, e.emp_code
, COUNT(e.dept_id) dept_id_count
, COUNT(e.sub_dept_id) sub_dept_id_count
FROM punching_data_table p
LEFT
JOIN employee_master e
ON e.emp_code = p.emp_code
GROUP
BY e.dept_id
, e.sub_dept_id
Any help on the subject with explanation will be highly appreciated.

Answer 1

This query (SQLFiddle) will give you the output that you specify:
SELECT em1.dept_id, em2.dept_count, em1.sub_dept_id, COUNT(em1.sub_dept_id) AS sub_dept_count
FROM employee_master em1
JOIN (SELECT dept_id, COUNT(dept_id) as dept_count
FROM employee_master e
JOIN punching_data_table p
ON e.emp_code = p.emp_code
GROUP BY dept_id) em2
ON em1.dept_id = em2.dept_id
WHERE EXISTS (SELECT * FROM punching_data_table pdt WHERE pdt.emp_code = em1.emp_code)
GROUP BY dept_id, sub_dept_id
Output:
dept_id dept_count sub_dept_id sub_dept_count
10 4 21 1
10 4 22 3
11 2 20 2
If you have a department names table e.g.
CREATE TABLE dept_master (dept_id INT, dept_name VARCHAR(20));
INSERT INTO dept_master VALUES
(10, 'Department 10'),
(11, 'Department 11');
You can join that as well using this query (SQLFiddle):
SELECT d.dept_name, em2.dept_count, em1.sub_dept_id, COUNT(em1.sub_dept_id) AS sub_dept_count
FROM employee_master em1
JOIN (SELECT dept_id, COUNT(dept_id) as dept_count
FROM employee_master e
JOIN punching_data_table p
ON e.emp_code = p.emp_code
GROUP BY dept_id) em2
ON em1.dept_id = em2.dept_id
JOIN dept_master d on d.dept_id = em1.dept_id
WHERE EXISTS (SELECT * FROM punching_data_table pdt WHERE pdt.emp_code = em1.emp_code)
GROUP BY d.dept_id, sub_dept_id
And then get department names in the output instead of ids:
dept_name dept_count sub_dept_id sub_dept_count
Department 10 4 21 1
Department 10 4 22 3
Department 11 2 20 2

Answer 2

Try this query
SELECT emp_code, MAX(dept_id_count) as dept_id_count, MAX(sub_dept_id_count) as sub_dept_id_count
FROM
(SELECT p.emp_code , COUNT(e.dept_id) dept_id_count, 0 as sub_dept_id_count
FROM punching_data_table p
LEFT JOIN employee_master e ON e.emp_code = p.emp_code
GROUP BY e.dept_id
UNION
SELECT p.emp_code, 0 as dept_id_count, COUNT(e.sub_dept_id) sub_dept_id_count
FROM punching_data_table p
LEFT JOIN employee_master e ON e.emp_code = p.emp_code
GROUP BY e.sub_dept_id) as subQuery
GROUP BY emp_code

Answer 3

There are some things which are not clear in your question for example do you want distinct employees or just a occurrences of employees, what happens if an employee does not exist in punching_data. I have made some assumptions in this answer but the crux of it is that you create 2 sub queries and join them on dept id which will force a 1 to many join
so given
drop table if exists employee_master,punching_data_table;
create table employee_master
(emp_code int, dept_id int, sub_dept_id int);
insert into employee_master values
(1 , 10 , 22),
(2 , 11 , 20),
(3 , 10 , 22),
(4 , 10 , 22),
(5 , 11 , 20),
(6 , 10 , 21 ),
(7 , 10 , 21);
create table punching_data_table(emp_code int);
insert into punching_data_table values
(1),
(2),
(3),
(4) ,
(5),
(6);
select a.dept_id,a.empcodes,s.sub_dept_id,s.empcodes from
(
select dept_id ,count(distinct emp_code) empcodes
from employee_master
where emp_code in (select emp_code from punching_data_table)
group by dept_id
) a
join
(select dept_id , sub_dept_id, count(distinct emp_code) empcodes
from employee_master
where emp_code in (select emp_code from punching_data_table)
group by dept_id,sub_dept_id) s
on s.dept_id = a.dept_id
order by a.dept_id,s.sub_dept_id
Result
+---------+----------+-------------+----------+
| dept_id | empcodes | sub_dept_id | empcodes |
+---------+----------+-------------+----------+
| 10 | 4 | 21 | 1 |
| 10 | 4 | 22 | 3 |
| 11 | 2 | 20 | 2 |
+---------+----------+-------------+----------+
3 rows in set (0.00 sec)

Answer 4

This result can be achieved without punching_data_table
SELECT t1.dept_id,
t2.sub_dept_id,
t1.count_dept,
t2.count_sub
FROM (SELECT dept_id,
sub_dept_id,
Count(dept_id) AS count_dept
FROM employee_master
GROUP BY dept_id)t1
LEFT JOIN (SELECT sub_dept_id,
dept_id,
Count(sub_dept_id) AS count_sub
FROM employee_master
GROUP BY sub_dept_id)t2
ON t1.dept_id = t2.dept_id