In our example, let us choose Employee as the entity name.

          <entity-group group="default" entity="Employee"/>

Here, the group attribute refers to the group name the delegator is associated with. You can find it in the WEB-INF/classes/entityengine.xml file, as shown in the following code snippet:

         <delegator name="default" entity-model-reader="main" 
           entity-group-reader="main">
           <group-map group-name="default" datasource-name="defaultDS"/>
         </delegator>

The entity attribute holds the name of the entity.

      <entity entity-name="Employee" table-name="employee" 
      package-name="">
        <field name="id" type="numeric"/>
        <field name="name" type="long-varchar"/>
        <field name="address" col-name="empaddress" 
         type="long-varchar"/>
        <field name="company" type="long-varchar"/>
        <prim-key field="id"/>
        <index name="emp_entity_name">
        <index-field name="name"/>
        </index>
      </entity>

Here, the entity-name attribute holds the name of the entity we have used in the previous step. The table-name attribute holds the name of the table; it is optional and will be derived from entity-name, if not present. package-name can be used if you want to organize and structure the entities' definitions into different packages.

The entity element contains one field element for each column in the table that needs to be created. The field element has a name attribute that holds the name of the field. If the column name of the field is different, the col-name attribute can be used, as in the case with employee address. If col-name is missing, the name of the field is used. The next important attribute is type. In our example, id is numeric whereas name and address are long-varchar.

These types of definitions of a field are mapped to the appropriate column type for each database type. The field-type mappings are stored under WEB-INF/classes/entitydefs/ and are declared in entityengine.xml, as shown in the following code snippet:

      <field-type name="h2" loader="maincp" 
      location="entitydefs/fieldtype-h2.xml"/>

If you look inside fieldtype-h2.xml, you will notice that numeric is mapped to BIGINT and long-varchar is mapped to VARCHAR. You can find out the various mappings and even the related Java data type from the same file.

The prim-key element is used to define the primarykey constraint for the table, as shown previously. In our case, id is the primary key. It is mandatory to name the primary key as id for all the new tables we are creating.

The index element creates a DB index for the field specified for that table. We can specify the index name and the group of the fields that needs to be indexed underneath it.

You can also define the relationship between entities using the element relation as shown in the following code snippet:

        <relation type="one"  rel-entity-name="Company">
              <key-map field-name="company" rel-field-name="id"/>
       </relation>

Here, we are adding a relationship between the Employee entity and the Company entity by saying that an employee can have only one company. In the preceding case, Employee should have a field company that points to the id field of a company's record. In other words, the company field in an employee's record will be the foreign key to the company's record.

More details on entity definitions can be found at https://cwiki.apache.org/confluence/display/OFBIZ/Entity+Engine+Guide#EntityEngineGuide-EntityModeling.