Question

I’m trying to run the following SQL statement:

IF OBJECT_ID('MyTable') IS NOT NULL
    DROP TABLE MyTable

SELECT
    a.UserId
INTO
    MyTable
FROM
    UsersTable a
WHERE
    a.UserId='12359670-1DC9-4A0A-8AE5-29B664C1A57E'


ALTER TABLE MyTable ALTER COLUMN UserId UNIQUEIDENTIFIER NOT NULL
ALTER TABLE MyTable ADD PRIMARY KEY(UserId) 

However, I get the following error:
Cannot define PRIMARY KEY constraint on nullable column in table ‘MyTable’.

Any ideas?

Answer 1

This assumes SQL Server based on UNIQUEIDENTIFIER

Put a GO between (or relevant batch separator if not SQL Server)

....
ALTER TABLE MyTable ALTER COLUMN UserId UNIQUEIDENTIFIER NOT NULL
GO
ALTER TABLE MyTable ADD PRIMARY KEY(UserId)

At batch compile time, the column is nullable. So break up the batches.

SQL isn’t a line by line procedural language

You’ll have to do this in a stored procedure

ALTER TABLE MyTable ALTER COLUMN UserId UNIQUEIDENTIFIER NOT NULL
EXEC('ALTER TABLE MyTable ADD PRIMARY KEY(UserId)')