Hi,

I've got two tables:

1) Image
2) ImageSpecification

Image.getSpecification() returns an object of type ImageSpecification

The problem I'm experiencing is that I see multiple ImageSpecification instances in the database with the exact same values, except for id. Since "id" is the primary key, the table allows this insertation.

My question is, how do I tell Hibernate: if you see another ImageSpecification in the table such that equals() returns true, use it instead of creating a new instance?

Thank you,
Gili

PS: I am aware I can explicitly look for duplicates in the table and use one if it exists. What I am asking for is a way to tell Hibernate to do this transparently.

Does this exist?

Gili

Uhm, no. Which one of your duplicates should Hibernate chose? We use the primary keys to distinquish tuples (as does your database) and if your primary keys are not the right way to identify a row, you should start thinking about your database content, esp. if its in a good and proper state or not.

Here is a hint: As a test of the integrity and correctness of your data, assume that "ImageSpecification" contains the size of the image in bytes. Now run an aggregate function, to find out how much bytes all of your images have. Getting the correct result?

P.S. "distinct" is not the right way to do it.

Christian,

I could use custom IDs and map id=hashCode() and that would likely fix the problem. The Hibernate documentation discourages the use of custom IDs, but would it not be appropriate in this case? As long as I ensure that hashCode() is unique (i.e. no collisions for ImageSpecifications with different values) it should be an appropriate solution, right?

Another problem is that updateOrSave() will no longer work, which is a bummer.

What do you think?

Thanks,
Gili

Really, stop thinking about Java and Hibernate, but think about your database! What use is duplicate information? Would you say it adds value or is it even possible that it is the source for false information?

I would say that duplicate information in the database is just a waste of space. I could allow duplication entries (no harm would be caused, it is not the source for false information) but I'd rather have at most one entry per possible ImageSpecification in the table.

In your example of ImageSpecification = image size, I would say I could scan all my images, get a list of all unique sizes and add them to the table. If a new image is added it uses a preexisting ImageSpecification if its size is already listed, otherwise it creates and adds a new one.

Gili

Well, if its only a waste of space, I'll rest my case. We have different opinions about relational databases and data management. ;)

Christian,

I don't understand. I know I've mentioned a specific use-case but I'm actually asking a theoretical question about Hibernate. Opinions about relational databases and data management should not matter.

On the topic of this specific case, can you please explain to me what your preferred course of action is? You mention a difference of opinion but I don't know what your opinion is :) Please enlighten me.

Thank you,
Gili

Clean up your broken database.

If we'd have a quote of the week, this would be it :) Questions about Hibernat can't possibly have something to do with relational databases. Wow...

You can

Christian,

Do you think the question I posted http://forum.hibernate.org/viewtopic.php?t=935398 is little bit related to what is described here

I also trying to avoid adding multiple columns in tables with exactly same data.

It will be great help if you can shed some more light on http://forum.hibernate.org/viewtopic.php?t=935398

Thanks in advance