Reverse Engineering "views" using Ant tasks

jai@hibernate · **Joined:** Thu Sep 07, 2006 4:35 am **Posts:** 9

We are using Hibernate 3 and are using the jdbcconfiguration task to autogenerate our hbm files and later on using hbm2java task in Ant to generate the java files. Things work fine when the hbms and java files are generated for "tables" in the schema. However, our schema even contains views and we generate the hbms and the java files using the same tools. The hbm/java files that are generated for these "views" look weird.

1) The hbm for the view has a id/composite-id defined for it

I am not great at databases, but as far as i know, views do not have something like an id. So, why are these ids generated?

2) If my view returns around 5 columns, then *all* these columns are included as part of the composite-id. Why is this?

max · **Posted:** Tue Jan 09, 2007 6:14 am

jai@hibernate wrote:

1) The hbm for the view has a id/composite-id defined for it

I am not great at databases, but as far as i know, views do not have something like an id. So, why are these ids generated?

hibernate still needs to know what identify an object/row thus it needs an id defined for the entities.

Quote:

2) If my view returns around 5 columns, then *all* these columns are included as part of the composite-id. Why is this?

see #1 answer.

jai@hibernate · **Joined:** Thu Sep 07, 2006 4:35 am **Posts:** 9

max wrote:

hibernate still needs to know what identify an object/row thus it needs an id defined for the entities.

Agreed.

max wrote:

2) If my view returns around 5 columns, then *all* these columns are included as part of the composite-id. Why is this?

see #1 answer.

But shouldn't there be a way through which we can say which columns need to be a part of the id(s). Currently, all the columns which dont make any sense being part of the id are getting included in the composite-id.

max · **Posted:** Tue Jan 09, 2007 6:57 am

jai@hibernate wrote:

But shouldn't there be a way through which we can say which columns need to be a part of the id(s). Currently, all the columns which dont make any sense being part of the id are getting included in the composite-id.

sure - did you try and read the docs ? :)

http://www.hibernate.org/hib_docs/tools ... ngineering

/max

jai@hibernate · **Joined:** Thu Sep 07, 2006 4:35 am **Posts:** 9

Hmmm.... I did read those docs, but did not pay much attention to the finer details :-) Blame my laziness for that ;-) Thanks for leading me to the docs. This is what we are talking about, right (Just to make sure i read the right configurations) :

Code:

<!-- table allows you to override/define how reverse engineering 
     are done for a specific table -->
<table name="ORDERS"> 
 <primary-key>
   <!-- setting up a specific id generator for a table -->
  <generator class="sequence">
    <param name="table">seq_table</param>
  </generator>
   <column name="CUSTID"/>
 </primary-key>
 <column name="NAME" property="orderName" type="string" />
 <!-- control many-to-one and set names for a specific named foreign key constraint -->
 <foreign-key constraint-name="ORDER_CUST">
   <many-to-one property="customer"/>
   <set property="orders"/>
 </foreign-key>
</table>

I will have to specify the id columns using the <primary-key> element above, isn't it?

max · **Posted:** Tue Jan 09, 2007 7:13 am

exactly

jai@hibernate · **Joined:** Thu Sep 07, 2006 4:35 am **Posts:** 9

Thanks Max for the help :-)