I have been using NetBeans 6.9.1 to create the POJO class and mapping files for the tables in the DB. There I can create both for tables with primary name but can't create for one without primary key. Is there any reason we should not or can't create POJO for table without primary key? There must be workaround for it am I right?

Anjib

Hello anjibman,

If there is no primary key, then you can create the pojo and hbm like this:

lets say there's a table Emp with 2 columns empid and ename both having not-null constraints and no primary key is there.

then you create Emp.java, EmpID.java and Emp.hbm.xml . EmpID.java will be having the properties empid and ename corresponding to the columns and Emp.java will be having a reference to the EmpID "eid"object. In the Emp.hbm.xml , you will be having a composite-id. The hbm looks similar to this:

<class name="com.Emp" table="Emp">
<composite-id name="eid" class="com.EmpID">
<key-property name="empid" column="empid"/>
<key-property name="ename" column="ename"/>
</composite-id>
</class>

using this you can work on the emp table which is not having a primary key. So, I think ur scenario is like this.

Prasad

My problem is I am using existing DB (Oracle). In this DB I have table called "USER" with fields user_id, name, email and phone but none of these fields are defined as primary key and I cannot alter the table schema.
Now my question is can I write mapping specifying user_id as <id> even its not a primary key in the DB?

Are there any other constraints like unique, not-null on those columns?

No constraints at all.
What difference will it make with or without constraints?

Now my question is can I write mapping specifying user_id as <id> even its not a primary key in the DB?
You can map it with an <id> element.
but it throws an exception if more than one record with a given id exists.
To explain it clearly,
Lets take the table u have mentioned: USER with USER_ID,NAME,EMAIL,PHONE as columns. If you create a POJO class by name User and have the properties as userId,name,email,phone and have the getters and setters. Then write the hbm file
with an id element with the property userId (generator assigned), and all other variables as normal properties.
Lets say there are 2 records with 101 as user_id. When you do a get like session.get(User.class,"101"); then hibernate throws an exception stating more than one row with the given identifier was found.

This bring to my next concern of generating incremental number on the field user_id. I think if I can generate incremental number in this field then exception can be avoided. Am I right?

Yes, If you made sure that the column which you have given as the id field has no duplicate values, then I think you are fine.

Just to make sure is that having

Also make sure that no external source source other than ur application is updating the table in the database. There might be a chance that another record with the same id can be entered as there is no unique constraint on the USER_ID field.