MySQL - count with coalesce and add missing rows - mysql

Question

I have one tables with two date columns (Date_open and Date_closed). All I want to do it to count occurrences per day. So to see how many were opened and closed each day. We look at the last 7 days from today. The problem is that some dates are not present in either of the columns and I can not find a way to either link tables with sub query (example code 1) or get coalesce work (example code 2)?
The table looks like that:
+------+------------+-------------+------+
| code | Date_open | Date_closed | Prio |
+------+------------+-------------+------+
| 1 | 2018-01-08 | 2018-01-08 | A |
| 2 | 2018-01-01 | 2018-01-08 | B |
| 3 | 2018-01-06 | 2018-01-07 | C |
| 4 | 2018-01-06 | 2018-01-06 | A |
| 5 | 2018-01-04 | 2018-01-06 | B |
| 6 | 2018-01-03 | 2018-01-01 | C |
| 7 | 2018-01-03 | 2018-01-02 | C |
| 8 | 2018-01-03 | 2018-01-02 | C |
+------+------------+-------------+------+
And the results I want are as follows:
Date OpenNo CloseNo
2018-01-01 1 1
2018-01-02 2
2018-01-03 3
2018-01-04 1
2018-01-05
2018-01-06 2 2
2018-01-07 1
2018-01-08 1 2
The first code I tried was:
SELECT *
FROM
(SELECT t1.Date_open,
COUNT(t1.Date_open) AS 'OpenNo'
FROM
Tbl AS t1
GROUP BY t1.Date_open)
AS A
JOIN
(SELECT t2.Date_closed,
COUNT(t2.Date_closed) AS 'CloseNo'
FROM
Tbl AS t2
GROUP BY t2.Date_closed)
AS B ON A.Date_open = B.Date_closed;
This code works as long as there is data for each day.
The second code I tried was:
SELECT
COALESCE (Date_open, Date_closed) AS Date1,
COUNT(Date_closed) AS ClosedNo,
COUNT(Date_open) AS OpenNo
FROM tbl
GROUP BY Date1;
Both do not work. Any ideas please?
Below is the code to create tbl.
create table Tbl(
code int(10) primary key,
Date_open DATE not null,
Date_closed DATE not null,
Prio varchar(10));
insert into Tbl values (1,'2018-01-08','2018-01-08' ,'A');
insert into Tbl values (2,'2018-01-01','2018-01-08' ,'B');
insert into Tbl values (3,'2018-01-06','2018-01-07' ,'C');
insert into Tbl values (4,'2018-01-06','2018-01-06' ,'A');
insert into Tbl values (5,'2018-01-04','2018-01-06' ,'B');
insert into Tbl values (6,'2018-01-03','2018-01-01' ,'C');
insert into Tbl values (7,'2018-01-03','2018-01-02' ,'C');
insert into Tbl values (8,'2018-01-03','2018-01-02' ,'C');

Answer 1

You may use a calendar table, and then left join to your current table twice to generate the counts for each date:
SELECT
d.dt,
COALESCE(t1.open_cnt, 0) AS OpenNo,
COALESCE(t2.closed_cnt, 0) AS CloseNo
FROM
(
SELECT '2018-01-01' AS dt UNION ALL
SELECT '2018-01-02' UNION ALL
SELECT '2018-01-03' UNION ALL
SELECT '2018-01-04' UNION ALL
SELECT '2018-01-05' UNION ALL
SELECT '2018-01-06' UNION ALL
SELECT '2018-01-07' UNION ALL
SELECT '2018-01-08'
) d
LEFT JOIN
(
SELECT Date_open, COUNT(*) AS open_cnt
FROM Tbl
GROUP BY Date_open
) t1
ON d.dt = t1.Date_open
LEFT JOIN
(
SELECT Date_closed, COUNT(*) AS closed_cnt
FROM Tbl
GROUP BY Date_closed
) t2
ON d.dt = t2.Date_closed
GROUP BY
d.dt
ORDER BY
d.dt;
Demo
The reason I aggregate the open and closed date counts in separate subqueries is that if were to try to just do a straight join across all tables involved, we would have to deal with double counting.
Edit:
If you wanted to just use the current date and seven days immediately preceding it, then here is a CTE which would do that:
WITH dates (
SELECT CURDATE() AS dt UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 1 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 2 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 3 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 4 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 5 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 6 DAY) UNION ALL
SELECT DATE_SUB(CURDATE(), INTERVAL 7 DAY)
)
You could inline the above into my original query which is aliased as d, and it should work.

Answer 2

Coalesce can be confusing - it returns the first non-null value from the list you provide to it.
I don't know if this question requires a super-complex answer.
To get the count of open and closed for each unique date, the following could work.
SELECT
COALESCE (Date_open, Date_closed) AS Date1,
SUM(IF(Date_closed != null,1,0)) AS ClosedNo,
SUM(IF(Date_open != null,1,0)) AS OpenNo
FROM tbl
GROUP BY Date1;