I'm trying to join vendor information into all product queries (read-only). For some reason, despite specifying the 'foreign-key' from our ITEM table, hibernate still generates code to join off of ITEM's primary key.

Am I missing something?

Hibernate version: 3.1

Mapping documents:


Name and version of the database you are using: Oracle 9i

The generated SQL (show_sql=true):

I think you want property-ref, not foreign-key.

I tried that too, but traced the issue to a clause in hibernate source that was commented with "always join by pkey"-- which seems to contradict the allowed mapping configuration

That sounds bad. You may have to switch to a one-to-one mapping, which is actually a many-to-one with unique="true" and property-ref specified (because one-to-one is hardwired to use the PK, too).

the foreign-key attribute would not achieve what you are trying; it is only used for explicitly naming the FK constraint during schema generation.

the property-ref attribute is what you want. if that is not working for you then that would be a bug.

Can you let me know how did you acheive this?

I have been trying to accomplish this...since a long time...

help is reallly appreciated.

I think its a bug; I am running into the same problem. I have a legacy database I need to join into using something other than the main objects primary key and it simply wont let me join on any field other than the primary key.

No matter what I insert into "property-ref" it ALWAYS uses the objects primary key in the inner join:




results in:



Where the inner join SHOULD be hperson0_.I_USER_LDAP =hperson0_1_.I_USER_LDAP. I_PERSN_IBSC is the primary key.

Is there a work around to this? If its a known problem has there been a patch jar released that I can deploy?

Thanks[/b]

It seems I cant use the <key /> tag to connect to another database table using another other than the primary key. The "property-ref" attributre doesnt seem to work under ANY scenario; I have tried creating a subclass to circumvent the join via foriegn key problem and also experience the same problem: I cannot connect into another table using anything other that the first tables primary key :(

Any suggestions? Is hibernate really that broken or am I doing something wrong?



<?xml version="1.0"?>
<!DOCTYPE hibernate-mapping PUBLIC
"-//Hibernate/Hibernate Mapping DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd">
<hibernate-mapping package="com.dcx.ps.gpsis.ibsc.data.entities" default-lazy="false">
<class
name="HPerson"
table="PIBPERSN">
<cache usage="read-write" />
<id name="id" column="I_PERSN_IBSC" type="long">
<generator class="increment"/>
</id>
<timestamp name="lastUpdated" column="T_STMP_UPD" />
<property name="lastUpdatedBy" type="string" column="I_USER_UPD" not-null="true" />
<property name="tid" column="I_USER_LDAP" type="string" not-null="true" unique="true" />
<property name="name" column="N_USER_LDAP" type="string" not-null="true"/>
<set name="positions" inverse="true" cascade="save-update">
<cache usage="read-write" />
<key column="I_PERSN_IBSC" />
<one-to-many class="HPosition" />
</set>
<!-- this is not functioning as the key tag seems to be broken in hibernate 3.x :(

<join table="PUSRPRF" inverse="true">
<key column="I_USER_LDAP" property-ref="tid" />
<property name="firstName" type="string" column="N_FIRST" />
<property name="lastName" type="string" column="N_LAST" />
</join>
-->
</class>
<!--
<query name="findPersonsByName">
<![CDATA[
from HProfilePerson person where trim(person.firstName) like :firstname% and trim(person.lastName) like :lastname%
]]>
</query>
<query name="findPersonsByFirstName">
<![CDATA[
from HProfilePerson person where trim(person.firstName) like :firstname%
]]>
</query>
<query name="findPersonsByLastName">
<![CDATA[
from HProfilePerson person where trim(person.lastName) like :lastname%
]]>
</query>
-->
<query name="findPersonByTid">
<![CDATA[
from HPerson person where trim(person.tid) = :tid
]]>
</query>
<query name="findPersons">
<![CDATA[
from HPerson
]]>
</query>
</hibernate-mapping>

sorry for the not format my mapping file; here it is again:

According to JIRA issue HHH-551, this is a known problem. Apparently there's a TODO for it.

Where are the TODOs listed? I'd have thought that they'd be JIRA issues, but Gavin closed 551, rather than using it as a tracking issue. I'm sure I've raised an issue or two that I wouldn't have, if I'd looked at the TODOs first.

I am curious

1. if there is a way to know what release this is planned. I got 3.1.2 and it isnt implemented.

2. another way to join from non PK column

Sincerely

MB

No, there's no way to find out when something is due, in an open source project. Not without directly asking, and that's likely to make enemies for you. After all, it's a labour of love for unpaid developers with other commitments.

I'd use a many-to-one unique="true" mapping and the delegation pattern to simulate a join via a non-PK column. Given TableA and TableB that you want to join into a single class PojoA, define a many-to-one unique="true" property-ref="foreign-key-column-in-A" mapping called something like "InternalLinkToB". Then delegate all the methods in PojoA to getInternalLinkToB().getXXX()/setXXX().

thanks for reply and for detailed suggestion, i'll do it that way.

as for my unfriendly question, all I wanted to say is I wish I'd knew what features are mentioned in docs as possible, but are not implemented.

of course, to all people responsible for this project big thank you. great work.

mb