Re[2]: [JBoss-user] Question about CMP/CMR

[Alex L. <lou...@ua...>]

Frank,

if I understood you right, you want to map a CMR field with foreign
key to a CMP field that is not a primary key. Currently, this is not
supported.

alex

Sunday, December 29, 2002, 3:32:33 PM, you wrote:

LF> Hallo Alex,

LF> I know you ported this feature not long ago from HEAD to 3.2.

LF> I use HEAD here and use this new feature for Bean2, if I understand it 
LF> right.
LF> Column festr.festa_serial is part of the primary key for festr and also 
LF> foreign key for festr referencing festa.festa_serial.

LF> But if the foreign key field is not part of the primary key, like in 
LF> Bean3 (febdr), the foreign key field must not be CMP-Field in the bean. 
LF> It's automatically added because of CMR.

LF> As said, I would like jboss to recognize that the field already exists 
LF> also in case of Bean3 and use this field for CMR.


LF> Frank


LF> Alex Loubyansky wrote:

>>Hello Frank,
>>
>>in JBoss-3.2 (cvs version) you can map foreign key fields to primary
>>key fields. This is done by specifying the same column names for
>>foreign and primary key fields.
>>
>>alex
>>
>>Saturday, December 28, 2002, 9:12:26 PM, you wrote:
>>
>>LF> JBoss-Head, JDK 1.4.1, Informix-DB
>>
>>LF> Following given situation:
>>LF> Bean 1:
>>LF>  Name=Festa, DBTable=festa, PK=festa_serial
>>LF> Bean 2:
>>LF>  Name=Festr, DBTable=festr, PK=festa_serial+festr_pos_nr, 
>>FK=festa_serial ->> festa
>>LF> Bean 3:
>>LF>  Name=febdr, DBTable=febdr, PK=febdr_serial, FK=festa_serial -> festa
>>
>>LF> As you can see, for bean 2 the CMR-Field festa_serial must be a 
>>LF> CMP-Field, because it's part of the PK.
>>LF> For bean 3 the CMR-Field festa_serial must not be a CMP-Field. Otherwise 
>>LF> you get a DeploymentException because of a "wrong" 
>>LF> SQL-CreateTable-Statement.
>>LF> The field festa_serial is mentioned twice here.
>>LF> I would like JBoss to recognize that the CMR-Field is already there as a 
>>LF> CMP-Field in both cases.
>>
>>LF> So my question: Is this a Bug, a FeatureRequest or a must be which will 
>>LF> not be changed ?


-- 
Best regards,
 Alex Loubyansky