Contraint names

gstamp · **Posted:** Mon Feb 09, 2004 2:47 am

I notice that (at least for SQL Server) the constraint names aren't exactly meaningful:

Code:

alter table SVD_CONTENTDOCUMENTCATEGORY add constraint FK1355E574C4195AD9 foreign key (CategoryId) references SVD_SEAVIEWCONTENT;

How are these names generated? Will they be consistent across runs of SchemaExport? From my experiments it seems to be the case but I'd like to know how the name is picked.

emmanuel · **Posted:** Mon Feb 09, 2004 6:04 am

Checking the sources will give you the fastest answer.

gstamp · **Posted:** Mon Feb 09, 2004 6:59 am

I guess I was being a bit lazy. The answer for anyone who's interested:

The FK names are calculated by taking a hash of table name and column names. Here's the code snippet:

Table.java:

Code:

   public String uniqueColumnString(Iterator iterator) {
      int result = 0;
      while ( iterator.hasNext() ) result += iterator.next().hashCode();
      return ( Integer.toHexString( name.hashCode() ) + Integer.toHexString(result) ).toUpperCase();
   }

chadmuco · **Posted:** Mon Nov 22, 2004 1:15 pm

In regards to the naming of constraints, is there a way to override the hash function and specify a constraint name in the hbm.xml file?

Thanks,
Chad

michael · **Posted:** Mon Nov 22, 2004 1:17 pm

We always thought about making this overideable by using a NamingStrategy - it was never implemented though :) Patches are welcome.

christian · **Posted:** Mon Nov 22, 2004 1:29 pm

There is a foreign-key="foo" mapping attribute, check the SchemaExport documentation chapter in the reference guide.