Unique colums automatically made NOT NULL

[Martin Dirichs <na...@dirichs.fastmail.fm>]

Hi,
when creating a database schema with OpenJPA, columns with a unique
constraint are automatically defined as non null. Even if the column is
explicitly anntotated as nullable (with @Column(nullable=true)), the
resulting table definition marks it as "NOT NULL".

Looking at the code it is not difficult to see why this happens. In
org.apache.openjpa.jdbc.schema.Unique it says:

    /**
     * Adds the given column. 
     * The added column is set to non-nullable because a unique constraint
     * on the database requires that its constituent columns are NOT NULL. 
     * @see Column#setNotNull(boolean)
     */
    public void addColumn(Column col) {
    	super.addColumn(col);
    	col.setNotNull(true);
    }


This is strange, since nullable unique columns are perfectly legal with
regard to ANSI SQL-92 as well as all databases I've work with so far, which
includes Oracle, MySQL, Postgresql and h2. In fact, this is a crucial
difference between columns defined as PRIMARY KEY and those only bound by a
UNIQUE constraint: the latter allows null values for the columns, unless
these have been explicitly defined as NOT NULL.

Is there a deeper reason for this behaviour of OpenJPA which I'm not aware
of?

-Martin.

-- 
View this message in context: http://n2.nabble.com/Unique-colums-automatically-made-NOT-NULL-tp2827809p2827809.html
Sent from the OpenJPA Users mailing list archive at Nabble.com.