Generating name based on ID and current year using SQL Server - sql

Question

I am trying to generate value for the ProjektNummer (Name) column based on the value of the ID (primary key column.
So for example, when ID is 142, then the ProjektNummer should be 19142. 19 indicating the current year and 142 is the value of its ID.
Now, when the year changes, the id part in the value of the project name column needs to be restarted from zero while the actual ID should follow the identity and be incremented by 1 as usual.
So if the last record in the Year 2019 has ID=164, the first record in the year 2020 should be:
ID: 165 ProjektNummer: 20001
.. and the second record in year 2020 would be:
ID: 166 ProjektNummer: 20002
One way to achieve this is by creating a new view say 'vMaxLastId' and storing the max value of 'ID' for its corresponding year. This record will be used as a reference in another trigger.
So, if the first record in the year 2020 has ID: 165 Another trigger will subtract the value (164) for year 2019 (which is stored in the view), from the value of 'ID' after inserting a new record in 'table1' using the second trigger. (this should be done before inserting tho)
165-164=1
166-164=2
167-164=3
....
When the year will change from 2019 to 2020 the max id and year in 2020 will be added as a new record in the view. I have implemented this but there are multiple problems associated with this approach.
Kindly suggest me any possible way to solve this problem. I am willing to make any kind of change to the DB, query or trigger or the whole concept.

Answer 1

The first thought that crossed my mind would be to create a SEQUENCE object to generate the ID portion of the value. Since there are no fancy rules around the numbering, just this should do.
CREATE SEQUENCE dbo.ProjektID;
That will give you a bigint. To consume it, CONCAT that to the current year, and put your padding zeros onto the sequence value. You could set this as the default value for the column.
CONCAT( YEAR( GETDATE() ) % 100 , RIGHT( CONCAT('000', NEXT VALUE FOR dbo.ProjektID), 3))
Then schedule a job to run at midnight on the first of the year to restart the sequence.
ALTER SEQUENCE dbo.ProjektID
RESTART;
I think that would cover your requirements.

Answer 2

The most pragmatic solution I could think of is the use of a INSTEAD OF trigger on that specific main table.
My test table:
CREATE TABLE dbo.MyTestProject
(
id INT IDENTITY(1, 1) NOT NULL PRIMARY KEY
, dateInserted DATETIME2(7) NOT NULL
, projectNumber INT NULL
);
And the example trigger:
CREATE TRIGGER dbo.triMyTestProject
ON dbo.MyTestProject
INSTEAD OF INSERT
AS
BEGIN
SET NOCOUNT ON;
DECLARE
#maxYear INT
, #maxNumber INT
, #startNumber INT
, #dateInserted DATETIME2;
SELECT
#dateInserted = dateInserted
FROM Inserted;
SET #startNumber = CAST(FORMAT(#dateInserted, 'yy') AS INT) * 1000 + 1;
SELECT
#maxNumber = MAX(projectNumber)
, #maxYear = ISNULL(MAX(YEAR(dateInserted)), YEAR(#dateInserted))
FROM dbo.MyTestProject;
INSERT INTO dbo.MyTestProject
(
dateInserted
, projectNumber
)
VALUES
(
#dateInserted, IIF(YEAR(#dateInserted) > #maxYear OR #maxNumber IS NULL, #startNumber, #maxNumber + 1)
);
END;
Add test data:
INSERT INTO dbo.MyTestProject (dateInserted) VALUES (SYSDATETIME());
INSERT INTO dbo.MyTestProject (dateInserted) VALUES (SYSDATETIME());
INSERT INTO dbo.MyTestProject (dateInserted) VALUES (SYSDATETIME());
INSERT INTO dbo.MyTestProject (dateInserted) VALUES ('20200102');
INSERT INTO dbo.MyTestProject (dateInserted) VALUES ('20200202');
Test result:
SELECT * FROM dbo.MyTestProject;
id dateInserted projectNumber
----------- --------------------------- -------------
1 2019-12-06 17:27:58.8813628 19001
2 2019-12-06 17:27:58.8833638 19002
3 2019-12-06 17:27:58.8883448 19003
4 2020-01-02 00:00:00.0000000 20001
5 2020-02-02 00:00:00.0000000 20002