Not sure if anyone cares, but Postgres is perfectly OK w/ me making FKs w/ the same name and the generator is not ...

PostrgreSQL 8.0.3-1 w/ postgresql-8.0-312.jdbc2.jar
version = hibernate-tools-3.0.0.alpha4a

org.hibernate.cfg.JDBCBinderException: Foreign key name (m_c_fk) mapped to different tables! previous: org.hibernate.mapping.Table(public.activity) current:org.hibernate.mapping.Table(public.meeting)
at org.hibernate.cfg.JDBCBinder.processForeignKeys(JDBCBinder.java:1062)
at org.hibernate.cfg.JDBCBinder.readDatabaseSchema(JDBCBinder.java:136)
at org.hibernate.cfg.JDBCBinder.readFromDatabase(JDBCBinder.java:91)
at org.hibernate.cfg.JDBCMetaDataConfiguration.readFromJDBC(JDBCMetaDataConfiguration.java:37)
at org.hibernate.eclipse.console.wizards.ArtifactGeneratorWizard$3.execute(ArtifactGeneratorWizard.java:237)
at org.hibernate.console.ConsoleConfiguration.execute(ConsoleConfiguration.java:77)
at org.hibernate.eclipse.console.wizards.ArtifactGeneratorWizard.buildConfiguration(ArtifactGeneratorWizard.java:234)
at org.hibernate.eclipse.console.wizards.ArtifactGeneratorWizard.doFinish(ArtifactGeneratorWizard.java:164)
at org.hibernate.eclipse.console.wizards.ArtifactGeneratorWizard.access$0(ArtifactGeneratorWizard.java:132)
at org.hibernate.eclipse.console.wizards.ArtifactGeneratorWizard$1.run(ArtifactGeneratorWizard.java:98)
at org.eclipse.jface.operation.ModalContext$ModalContextThread.run(ModalContext.java:113)

hmmm....most db's complain about that AFAIK....

I had the exact same problem after importing my database into postgres and all the foreign keys were set to "$1", "$2", etc...

Couldn't find any edit button.. so..

org.hibernate.cfg.JDBCBinderException:Foreign key name ($1) mapped to different tables! previous: org.hibernate.mapping.Table(public.tcontactupdater) current:org.hibernate.mapping.Table(public.temroxcontact)

That was the error report I got while using artifact generation plugin for eclipse...

Postgres creates foreign keys named "$1", "$2", etc by default if you use this kind of syntax

CREATE TABLE order (
orderid serial PRIMARY KEY NOT NULL,
id_contactid int4 REFERENCES tcontact
)

the "REFERENCES tcontact" part is the same as doing:

ALTER TABLE torder
ADD CONSTRAINT "$1" FOREIGN KEY (id_contactid) REFERENCES tcontact (contactid) ON UPDATE NO ACTION ON DELETE NO ACTION;

if contactid is the primary key of tcontact.

In other words, if you are using Postgres and the REFERENCES syntax you can't use the hibernate tools without dropping all your foreign keys and creating them with names...

So... is there any chance of non-unique keys beeing allowed in the hibernate tools for generating xml/java?

Add this explanation to JIRA and lets see.

In other db's the foreign key name is unique (and actually also inside hibernate in some aspects); but we could probably handle it with a bit of work....patches welcome....

but i would recommend you name those foreign keys since otherwise the names would be pretty bad.

according to postgres docs that is done most compact via:

id_contactid int4 REFERENCES myFkName tcontact

(dont have a postgres install so let me know if that is actually possible)

Hi, I had looked for what you just described, but couldn't find it... see documentation below.

I tried it as well, but it didn't work. (This is on postgres 7.4, I can check on postgres 8.0.1 for windows later)

Personally I don't do anything with foreign keys after they have been created and therefore haven't thought about nameing them in any useful fashion... and Postgres really doesn't care...

I haven't looked at the implementation of the tools, but you could generate unique names by using tablename+"_"+columname...

Seems like I am gonna drop "unnamed"/create "named" foreign keys in my database... Thanks for the reply, you probably get better support here than in commercial products :)

Name
CREATE TABLE -- define a new table
Synopsis
CREATE [ [ LOCAL ] { TEMPORARY | TEMP } ] TABLE table_name (
{ column_name data_type [ DEFAULT default_expr ] [ column_constraint [, ... ] ]
| table_constraint } [, ... ]
)
[ INHERITS ( parent_table [, ... ] ) ]
[ WITH OIDS | WITHOUT OIDS ]

where column_constraint is:

[ CONSTRAINT constraint_name ]
{ NOT NULL | NULL | UNIQUE | PRIMARY KEY |
CHECK (expression) |
REFERENCES reftable [ ( refcolumn ) ] [ MATCH FULL | MATCH PARTIAL ]
[ ON DELETE action ] [ ON UPDATE action ] }
[ DEFERRABLE | NOT DEFERRABLE ] [ INITIALLY DEFERRED | INITIALLY IMMEDIATE ]

and table_constraint is:

[ CONSTRAINT constraint_name ]
{ UNIQUE ( column_name [, ... ] ) |
PRIMARY KEY ( column_name [, ... ] ) |
CHECK ( expression ) |
FOREIGN KEY ( column_name [, ... ] ) REFERENCES reftable [ ( refcolumn [, ... ] ) ]
[ MATCH FULL | MATCH PARTIAL ] [ ON DELETE action ] [ ON UPDATE action ] }
[ DEFERRABLE | NOT DEFERRABLE ] [ INITIALLY DEFERRED | INITIALLY IMMEDIATE ]

I read it on http://www.postgresql.org/docs/8.0/inte ... TRAINTS-FK

and they are not very clear on how you actually name stuff when using REFERENCES.

Besides:

"So, to specify a named constraint, use the key word CONSTRAINT followed by an identifier followed by the constraint definition. (If you don't specify a constraint name in this way, the system chooses a name for you.)"

and then later on

"You can assign your own name for a foreign key constraint, in the usual way."

so maybe they mean:

product_no integer CONSTRAINT your_name REFERENCES products,

I did some more research, and postgres 8.0 for windows creates foreign keys like this: torder_id_modelid_fkey if you use REFERENCES. I will just go ahead and rename my stupidly named foreign keys then ;)

so postgres 8 looks much better ;)

Did you possibly find any docs about why they changed to this better strategy ?

No, I just double checked in my 8.0 db, as it sounded idiotic to create names like $1. The postgres developers must have seen it as well, because its fixed ;)

Luckily I haven't been using the "REFERENCES" syntax for too long, so I only have about 100-200 foreign keys to fix.

What fun.