Primary and Unique Key both enforce uniqueness of columns on which they are defined. Then where do they differ? They differ in following way
– A primary key doesn’t allow null value wherein a unique key allows one null value.
– A primary keys purpose is to uniquely define a row in a table wherein unique keys purpose is to enforce uniqueness on a column(s).
– A primary key is implemented by a default unique clustered index on column(s) wherein unique key is implemented by a default unique non clustered index on column(s).
– A primary key can’t be created on a null able column wherein a unique key can be created on a null able column.
– There can be only one primary key on a table wherein a table can have multiple unique key defined on it.
Let’s create a table with a primary key and a unique key constraint.
CREATE TABLE tblCns (sno INT NOT NULL, col1 VARCHAR(100),col2 VARCHAR(100)) GO -- Add unique key constraint ALTER TABLE tblCns ADD CONSTRAINT uq_Cns_col1 UNIQUE(col1) GO -- Add another unique constraint ALTER TABLE tblCns ADD CONSTRAINT uq_Cns_col2 UNIQUE(col2) GO -- Add a primary key constraint ALTER TABLE tblCns ADD CONSTRAINT pk_Cns_sno PRIMARY KEY(sno) GO
Let’s analyze the indexes using sp_helpindex stored procedure.
There are two unique key and a primary key. Attempt to create a second primary key will result in an error. The primary key is implement by a unique clustered index and unique key is implemented by a unique non clustered index.