[barrina]

Hi,

I am using Hibernate 3 within MyEclipse 4.1.1. I am new to hibernate and have question about why the MyEclipse is generating all the columns in my table within a composite-id tag and adding a secondary ‘Id’ POJO. I have read a related thread http://www.myeclipseide.com/PNphpBB2+file-viewtopic-t-11684-highlight-compositeid.html, that comes close, but doesn’t really answer my problem.

Under what circumstances will MyEclipse wrap the tables columns in a composite-id tag? I read previously that it was when the number of columns making the primary key was greater than zero. The table I have entered below as way of an example has 3/5 column that are not null, but no primary key — does MyEclipse infer that as the composite key? What about the remaining two coumns, should they be outside of the compostie-id tag and within the normal POJO rather than the POJOId class.

This all seems a bit of a black art — any clarification would be greatly appreciated.

=====
Table creation script
=====

      
      /*
      ACTION is CREATE Table gl
      */

CREATE TABLE gl (
       facility_id          NUMBER(10) NOT NULL,
       dept_code            VARCHAR2(15) NOT NULL,
       gl_account           VARCHAR2(20) NOT NULL,
       gl_category          NUMBER NULL,
       descr                VARCHAR2(50) NULL
);

There is no primary key or index defined.

====
Generated *.hbm.xml
====

<?xml version="1.0"?>
<!DOCTYPE hibernate-mapping PUBLIC "-//Hibernate/Hibernate Mapping DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">
<!-- 
    Mapping file autogenerated by MyEclipse - Hibernate Tools
-->
<hibernate-mapping>
    <class name="com.premierinc.informatics.oa.pr.pojo.Gl" table="GL" schema="OAR">
        <composite-id name="id" class="com.premierinc.informatics.oa.pr.pojo.GlId">
            <key-property name="facilityId" type="java.lang.Long">
                <column name="FACILITY_ID" precision="10" scale="0" />
            </key-property>
            <key-property name="deptCode" type="java.lang.String">
                <column name="DEPT_CODE" length="15" />
            </key-property>
            <key-property name="glAccount" type="java.lang.String">
                <column name="GL_ACCOUNT" length="20" />
            </key-property>
            <key-property name="glCategory" type="java.lang.Long">
                <column name="GL_CATEGORY" precision="22" scale="0" />
            </key-property>
            <key-property name="descr" type="java.lang.String">
                <column name="DESCR" length="50" />
            </key-property>
        </composite-id>
    </class>
</hibernate-mapping>

And I have a Gl.java and GlId.java — can be posted if required.

Kind regards,

Alan Barrington-Hughes

====

[Riyad Kalla]

Every table Hibernate maps has to have a PK. If it doesn’t, Hibernate will make one by using every field in the table. Just give your table a PK and Hibernate will stop going nuts making one for you 😉

[yxan]

I found this issue and would like to add to this, it would be nice if the plug could allow you to ignore doing that. I am planning to do this as a Select only so a PK is not of any greater concern to me.

so can you create a Hibernate mapping file with no PK? if you only use Selects or will it still blow up?

[Riyad Kalla]

It will still blow up, try and remove the ID section from the hbm.xml file, it’s invalid according to the Hibernate spec file. Hibernat needs it.

[Author]

Posts