Question

I’m currently using

SELECT MAX(CustomerID)...

to select the CustomerID, which is the primary key of the table, number of a newly added row because I need to use this number for another table that uses it as a foreign key. One row is added via INSERT, and one SELECT query is performed sequentially.

Under column property in SQL Server Management Studio, I see the Identity Increment is set to 1. Is there an unsafe reason to use MAX(CustomerID) to get what I need? The code will always be “one row INSERT and one SELECT queries” for this procedure. If there is a better way, what would that be? Perhaps, somewhere along the line, “get me the CustomerID of the newly added row”?

Answer 1

Your suggested approach with MAX(CustomerID) will fail under conditions of concurrency (and when the table is empty)

use SELECT SCOPE_IDENTITY() or the OUTPUT clause. Examples of both below

CREATE TABLE #Customers
(
CustomerID INT IDENTITY(1,1) PRIMARY KEY,
Name VARCHAR(50)
)

INSERT INTO #Customers
(Name) 
VALUES ('Customer 1')

SELECT SCOPE_IDENTITY()

INSERT INTO #Customers
(Name) 
OUTPUT inserted.CustomerID
VALUES ('Customer 1')


DROP TABLE #Customers