Hello!

When I abstract from Hibernate using the DAO Pattern, how can I avoid a lost update? For example I execute the following steps:

1. I call a "read" DAO method, which uses Hibernate to load the data I need from the database and gives me back a DTO.
2. I do something with this data and modify it.
3. I write it back by calling an "update" DAO method, which again uses Hibernate to make the changes persistent.

If I have two users, which execute these steps in parallel, I could end up with a lost update. One solution would be to implement my own optimistic locking, but isn't there a better way where I can use the optimistic locking of Hibernate? Or a different pattern?

Thank you very much for any suggestions or links,
Jonas

P.S.: I posted a similar question on theserverside, but I thought maybe someone here had the same problem.

Map a version column and use Hibernate's automatic versioning (or better, in your case, use the same database transaction for both DAO calls if there is no "long" time in between). Check the documentation, chapter about Transactions.