MSSQL - Counting dates from two columns grouping by external date range - sql-server

Question

I have a simple table containing case number (ID), opening_date and end_date where end_date have null values (unfinished cases). It looks like this:
ID opening_date end_date
1 2021-01-04 2021-01-14
2 2021-01-04 2021-01-26
3 2021-01-14 2021-02-15
4 2021-02-01 NULL
5 2021-02-04 2021-02-26
6 2021-02-10 2021-02-15
I'm trying to write a select query which will show me simply by month/week or day (nevermind), how many cases were set up (opening_date) and how many were closed (end_date) per each month./week. The problem is that I cannot use opening or end date in filters because not every date from opening_date column is in end_date and vice versa. It should be specific date range generated separately as external table in first column or something like that, so if there's situation, where neither opening nor end date occurs (in a day/week/month), a row with zeros should be shown as in the first date below - The result of example by day:
date openings endings
2021-01-01 0 0
2021-01-02 0 0
2021-01-03 0 0
2021-01-04 2 0
2021-01-05 0 0
2021-01-06 0 0
2021-01-07 0 0
2021-01-08 0 0
2021-01-09 0 0
2021-01-10 0 0
2021-01-11 0 0
2021-01-12 0 0
2021-01-13 0 0
2021-01-14 1 1
2021-01-15 0 0
2021-01-16 0 0
2021-01-17 0 0
2021-01-18 0 0
2021-01-19 0 0
2021-01-20 0 0
2021-01-21 0 0
2021-01-22 0 0
2021-01-23 0 0
2021-01-24 0 0
2021-01-25 0 0
2021-01-26 0 1
2021-01-27 0 0
2021-01-28 0 0
2021-01-29 0 0
2021-01-30 0 0
2021-01-31 0 0
2021-02-01 1 0
2021-02-02 0 0
2021-02-03 0 0
2021-02-04 1 0
2021-02-05 0 0
2021-02-06 0 0
2021-02-07 0 0
2021-02-08 0 0
2021-02-09 0 0
2021-02-10 1 0
2021-02-11 0 0
2021-02-12 0 0
2021-02-13 0 0
2021-02-14 0 0
2021-02-15 0 2
2021-02-16 0 0
2021-02-17 0 0
2021-02-18 0 0
2021-02-19 0 0
2021-02-20 0 0
2021-02-21 0 0
2021-02-22 0 0
2021-02-23 0 0
2021-02-24 0 0
2021-02-25 0 0
2021-02-26 0 1
2021-02-27 0 0
2021-02-28 0 0
By months:
Month openings endings
2021-01 3 2
2021-02 3 3
Please help me. Thanks in advance.

Answer 1

You need a calendar table for this. You start with the calendar, and LEFT JOIN everything else.
To get the calculation for each day, we can unpivot and group, then count daily totals
You can have a real table. Or you can generate it on the fly, like this:
WITH
L0 AS ( SELECT c = 1
FROM (VALUES(1),(1),(1),(1),(1),(1),(1),(1),
(1),(1),(1),(1),(1),(1),(1),(1)) AS D(c) ),
L1 AS ( SELECT c = 1 FROM L0 A, L0 B, L0 C ),
Nums AS ( SELECT rownum = ROW_NUMBER() OVER(ORDER BY (SELECT 1))
FROM L1 ),
Dates AS ( SELECT [date] = DATEADD(day, rownum, '20180101')
FROM Nums )
SELECT
d.[date],
openings = ISNULL(t.openings, 0),
endings = ISNULL(t.endings, 0)
FROM Dates d
LEFT JOIN (
SELECT v.AllDates,
openings = COUNT(IsOpen),
endings = COUNT(IsEnd)
FROM YourTable t
CROSS APPLY (VALUES
(opening_date, 1, NULL),
(end_date, NULL, 1)
) v(AllDates, IsOpen, IsEnd)
GROUP BY v.AllDates
) t ON t.AllDates = d.[date];