In this post, I am sharing a demonstration on how to allow multiple NULLs for UNIQUE Constraint in SQL Server.

As per the ANSI, UNIQUE constraint allows multiple NULLs. But in the SQL Server, it allows only one NULL value.
With the UNIQUE constraint, you cannot insert multiple NULLs.

But you can create UNIQUE NONCLUSTERED INDEX with the NOT NULL filter and can insert multiple NULLs.

Please check the below demonstration:

Create a sample table with UNIQUE Constraint:

CREATE TABLE tbl_TestUnique (ID INT UNIQUE)
GO

Try to insert multiple NULLs:

INSERT INTO tbl_TestUnique 
VALUES (1),(NULL),(NULL),(2),(NULL)
GO


Msg 2627, Level 14, State 1, Line 4
Violation of UNIQUE KEY constraint 'UQ__tbl_Test__3214EC26DBD9CD71'. Cannot insert duplicate key in object 'dbo.tbl_TestUnique'. The duplicate key value is ().

Now, try to insert with single NULL:

INSERT INTO tbl_TestUnique 
VALUES (1),(2),(NULL)
GO

Drop the table:

DROP TABLE tbl_TestUnique
GO

Create the same table with UNIQUE NONCLUSTERED INDEX:

CREATE TABLE tbl_TestUnique (ID INT)
GO

CREATE UNIQUE NONCLUSTERED INDEX idx_tbl_TestUnique_ID
ON tbl_TestUnique(ID)
WHERE ID IS NOT NULL
GO

Now, INSERT Multiple Nulls:
As we defined UNIQUE on ID NOT NULL, so we can insert multiple NULLs.

INSERT INTO tbl_TestUnique 
VALUES (1),(2),(NULL),(NULL)
GO

INSERT INTO tbl_TestUnique 
VALUES (NULL),(NULL),(NULL)
GO

In this post, I am sharing a solution on how to allow only single NULL record in the UNIQUE Constraint Column of PostgreSQL.

PostgreSQL: Composite UNIQUE Constraint does not consider NULL Values

Once you define the UNIQUE constraint, you can insert N number of NULL values for that column which is the principal rule of UNIQUE Constraint.

What if I need an only single NULL record for UNIQUE Column, we should check the below demonstration.

Create a sample table with UNIQUE Constraint:

CREATE TABLE tbl_testunique (ID INTEGER UNIQUE);

Insert sample multiple NULLs:
You can insert all NULLs.

INSERT INTO tbl_testunique 
VALUES (1),(NULL),(NULL),(NULL);

Now, Truncate the table:

INSERT INTO tbl_testunique 
VALUES (1),(NULL),(NULL),(NULL);

Create a UNIQUE INDEX with IS NULL:

CREATE UNIQUE INDEX idx_id ON tbl_testunique ((ID IS NULL)) 
WHERE ID IS NULL;

Now, Try to execute same INSERT statement:

INSERT INTO tbl_testunique 
VALUES (1),(NULL),(NULL),(NULL);

Will get ERROR in the result:

********** Error **********
ERROR: duplicate key value violates unique constraint "idx_id"
SQL state: 23505
Detail: Key ((id IS NULL))=(t) already exists.

I am sharing this interview question for Database Developer who has around 1 or 2 years of experience.

Many times, I have asked the difference between Unique Constraint and Unique Index but most of the time people answers were like “No difference – Both are same”. This is a wrong answer.

Which one is the better:

When you are creating a column using UNIQUE constraint or UNIQUE Index, in both the cases SQL Server by default creates a unique non-clustered index.
There is no any performance difference between these two, so you can use any one of this.
But still, there are few minor differences between Unique Constraint and Unique Index.

Disable:

Unique Constraint: You cannot disable the Unique Constraint, you will get below error:
Msg 4916, Level 16, State 0, Line 1
Could not enable or disable the constraint. See previous errors.

Unique Index: You can disable it

ALTER INDEX Index_name ON Table_name DISABLE

Errors: The error messages for both violations are also different

Unique Constraint violation error:

Msg 2627, Level 14, State 1, Line 1
Violation of UNIQUE KEY constraint 'uk_tb'. Cannot insert duplicate key in object 'dbo.mytab'. The duplicate key value is (1).
The statement has been terminated.

Unique Index violation error:

Msg 2601, Level 14, State 1, Line 1
Cannot insert duplicate key row in object 'dbo.col1' with unique index 'id_tb'. The duplicate key value is (1).
The statement has been terminated.

Filters:

Unique Constraint: You cannot add filter in Unique Constraint.
Unique Index: You can add filter in Unique Index:

CREATE UNIQUE NONCLUSTERED INDEX idx_name 
ON dbo.table_name(column_name) WHERE column_name ='M'

The Data validation is always a major concern of any Database System.
We all know about the UNIQUE constraint which prevents the duplicate records.

One of the Previous article, I have already shared one small note about the UNIQUE Index of PostgreSQL.

PostgreSQL: Composite UNIQUE Constraint does not consider NULL Values

In this post, I am sharing a similar post for Partial Unique Index of PostgreSQL.

For example, I want to stop duplicate Student Grages for particular Student Division only.

Create a sample table:

CREATE TABLE tbl_Students
(
	StudID INT 
	,StudName CHARACTER VARYING 
	,StudDivision CHARACTER VARYING
	,StudGrade CHAR(1)
);

Create a Partial UNIQUE Index:

CREATE UNIQUE INDEX tbl_Students_StudGrade ON tbl_Students (COALESCE(StudDivision,''),COALESCE(StudGrade,'')) 
WHERE StudDivision IN ('First','Second','Third');

Insert few sample records:
As we have creted Partial index so we can insert only unique combination of StudDivision (First,Second,Third) and StudGrade.

INSERT INTO tbl_Students
VALUES 
(1,'Anvesh','First','A')
,(2,'Jeeny','Second','B')
,(3,'Roy','Third','B')
,(4,'Mukesh','First','C')
,(5,'Martin','Second','C')
,(6,'Kavita','Third','A');

Try to insert few Invalid records:

INSERT INTO tbl_Students
VALUES 
(7,'Rajesh','First','A')
,(8,'Nile','Second','B');

Try to insert few records which are not validated by Partition Index:
We can insert duplicate records StudGrade + StudDivision for other StudDivisions except Partial StudDivision.

INSERT INTO tbl_Students
VALUES 
(7,'Dimpi','Fourth','C')
,(8,'Bony','Fourth','C');