Get dates missing from multiple date ranges - sql

Question

I have one table that stores when a customer support employee is in a particular location and for what date. Each separate date is its own record.
I have a second table that stores a range of dates that customers have asked for onsite support.
I need to extract a list of dates that a given location does NOT have any support representation. All I need is the location and the date(s). I don't care which employee in that location or which customer has requested the support.
So in the sample data below, I need to see as my query results:
+--------+------------+
| London | 04/01/2019 |
| London | 07/01/2019 |
| Paris | 05/01/2019 |
+--------+------------+
Table: Employee_Location
+----------+----------+------------+
| Employee | Location | Date |
+----------+----------+------------+
| 1111 | London | 01/01/2019 |
| 1111 | London | 02/01/2019 |
| 1111 | London | 03/01/2019 |
| 2222 | Paris | 01/01/2019 |
| 2222 | Paris | 02/01/2019 |
| 2222 | Paris | 03/01/2019 |
| 2222 | Paris | 04/01/2019 |
| 3333 | London | 05/01/2019 |
| 3333 | Paris | 06/01/2019 |
| 3333 | Paris | 07/01/2019 |
| 4444 | London | 06/01/2019 |
+----------+----------+------------+
Table: Customer_Request
+----------+----------+---------------+------------+
| Customer | Location | Request From | Request To |
+----------+----------+---------------+------------+
| AAAA | London | 01/01/2019 | 06/01/2019 |
| BBBB | Paris | 01/01/2019 | 06/01/2019 |
| CCCC | London | 05/01/2019 | 07/01/2019 |
+----------+----------+---------------+------------+
Here is my current code ...
select c.CALENDARDTM
from CALENDAR c, Employee_Location el
join Customer_Request cron el.location = cr.location
where c.CALENDARDTM NOT BETWEEN cr.RequestFrom and cr.RequestTo
and c.CALENDARDTM between '2019-01-01' AND '2019-01-07'

Answer 1

The key to solving this problem is to create a recordset that contains all dates between your nominated start and end dates.
There are a variety of methods you can use to do this, in the below example I have used a recursive CTE, for larger datasets you will need to tweak this slightly.
Once you have a list of all the dates, you combine it with a list of all locations, so you have all dates at all locations.?
Then you remove all records which match the records that already exists, in the example below a 'Not Exists' is used, but you can use a variety of approaches to get the desired outcome.
CREATE TABLE #Employee_Location (Employee int, [Location] varchar(100), [date] date)
INSERT INTO #Employee_Location (Employee, [Location], [Date])
VALUES (1111,'London','2019-01-01')
,(1111,'London','2019-01-02')
,(1111,'London','2019-01-03')
,(2222,'Paris','2019-01-01')
,(2222,'Paris','2019-01-02')
,(2222,'Paris','2019-01-03')
,(2222 ,'Paris','2019-01-04')
,(3333,'London','2019-01-05')
,(3333,'Paris','2019-01-06')
,(3333,'Paris','2019-01-07')
,(4444,'London','2019-01-06')
DECLARE #StartDate date = '2019-01-01'
DECLARE #EndDate date = '2019-01-07'
;WITH Dates AS (
SELECT #StartDate as d
UNION ALL
SELECT DateAdd(d, 1, d) as d
FROM Dates
WHERE d < #EndDate
)
,Locations AS (
SELECT DISTINCT [Location]
FROM #Employee_Location
)
,AllRecords AS (
SELECT d
,[Location]
FROM Dates
FULL OUTER JOIN Locations
ON 1=1
)
SELECT *
FROM AllRecords
WHERE NOT EXISTS (SELECT 1
FROM #Employee_Location e
WHERE e.[date] = Allrecords.d
AND e.[Location] = Allrecords.[Location])