Creating Modified Rows from Frequency Column in T-SQL - sql

Question

I need some ideas regarding an efficient way of creating rows per each count on a frequency column on SQL. (SQL Server 2016)
The data:
I have a table with the dates people called in sick and how many days they said they were gonna be absent:
BEGIN_DATE DAYS_SICK
2011-01-01 00:00:00.000 3
2011-01-01 00:00:00.000 3
2011-01-01 00:00:00.000 1
2011-01-02 00:00:00.000 2
2011-01-02 00:00:00.000 3
2011-01-04 00:00:00.000 4
2011-01-04 00:00:00.000 4
2011-01-04 00:00:00.000 3
I want to translate this to a table where each row represents a day in the year and I count the number of people that are sick that day.
DATE PEOPLE_SICK
2011-01-01 00:00:00.000 3
2011-01-02 00:00:00.000 4
2011-01-03 00:00:00.000 4
2011-01-04 00:00:00.000 4
2011-01-05 00:00:00.000 3
2011-01-06 00:00:00.000 3
2011-01-07 00:00:00.000 2
So for example:
For 2011-01-01 there were 3 persons that called in sick, 2 called in sick for 3 days and one only for that day. The output is 3.
Now on 2011-01-02 another 2 (different) persons called in sick but there were 2 persons from the day before that said they were gonna miss that day too so the output is 4.
No person called in sick on 2011-01-03 but there were 2 persons from 2 days ago that said they were gonna miss that day plus 2 persons from the day before. The output is 4.
Etc...
I am currently doing this by iterating through each of the rows in the input and then looping over the frequencies, adding or updating rows on the new table as necessary but it takes an obscene amount of time.
Is there any other way of doing this more efficiently?

Answer 1

This doesn't deal with weekends at all but can get you started. Also if there were a query that ran often I would build a DATE DIM table and use it instead of the Dates CTE. Where I got the DATE DIM code from.
CREATE TABLE #test (ID int IDENTITY(1,1), BEGIN_DATE datetime, DAYS_SICK int);
DECLARE #StartDate datetime = '2011-01-01'
, #CutoffDate datetime = '2011-01-10';
INSERT INTO #test (BEGIN_DATE, DAYS_SICK)
VALUES
('2011-01-01 00:00:00.000', 3),
('2011-01-01 00:00:00.000', 3),
('2011-01-01 00:00:00.000', 1),
('2011-01-02 00:00:00.000', 2),
('2011-01-02 00:00:00.000', 3),
('2011-01-04 00:00:00.000', 4),
('2011-01-04 00:00:00.000', 4),
('2011-01-04 00:00:00.000', 3);
WITH Dates
AS (SELECT d
FROM (
SELECT d = DATEADD(DAY, rn - 1, #StartDate)
FROM (SELECT TOP (DATEDIFF(DAY, #StartDate, #CutoffDate)) rn = ROW_NUMBER() OVER (
ORDER BY s1.[object_id])
FROM sys.all_objects AS s1
CROSS JOIN sys.all_objects AS s2
ORDER BY s1.[object_id]
) AS x
) AS y
)
,SickRanges
AS (
SELECT BEGIN_DATE
,DATEADD(DAY, DAYS_SICK - 1, BEGIN_DATE) END_DATE
FROM #test
)
SELECT d.d [DATE]
,count(1) PEOPLE_SICK
FROM SickRanges sr
JOIN Dates d ON d.d BETWEEN sr.BEGIN_DATE AND sr.END_DATE
GROUP BY d.d
ORDER BY d.d
DROP TABLE #test