Hibernate version: 3.1

Hi. I am having a problem figuring out how to have Hibernate enforce uniqueness on a set of columns. Basically, we have a legacy database that contains four(!) primary keys (the reason there are four is because other applications need it). From our application's perspective, we want to define uniqueness based on two of those columns. Now if one attempts to insert a row with the same values for those two columns we would like for Hibernate to stop it. The DBMS won't do it beacuse it doesn't violate any constraints.

I have tried doing this using composite-id and it allows for duplicate entries. I also tried using natural-id and properties (in both cases, I defined a separate surrogate key - identity column) but they all allowed duplicate entries.

This is a simple scenario of one table with no associations.

Thanks you in advance.

Isaac
[/code]

Hi,

Have you tried attribute unique-key="multicolumn_unique_key_name" in the mappings.

Check this out.

From my understanding of the documentation , as well as the attribute of , is useful for DDL generation. I have legacy tables. [/code]

I do not think hibernate does any data validation checks. So you will have to do this in your application yourself.

You may be right. However I would argue that Hibernate should provide this because they are closer to the database. Imagine that I do a validate check and it finds that there are indeed no duplicate records (as defined by me) out there. It is possible for another user to now sneak in a record (before the actual save) thereby violating the data integrity.

Even if hibernate did this, your scenario is not possible to avoid without doing something on the database.

You are correct. I guess I'll need to lock the table. I was just hoping that Hibernate would shield me from this.