Hi,

I have the following requirement:

1. Select a particular row from a table..
2. If the row is not found, insert a new one.

Now if two processes try to do the same concurrently, one of them will get a unique constraint (primary key) violation on step 2. Both processes will try to insert the same exact row, so there is no data correctness issue.

This looks like it might be a pretty standard problem. Is there there a "standard" solution for this in hibernate.

Thanks,
-Gautam

For me the 'standard solution' for this standard problem is to
properly catch and handle the unique constraint violation exception in your implementation.

1. lookup a row from a table with a determinate key.
2. If the row is not found, try to insert a new one.
3. If constraint violation exception raises, rollback, begin a new transaction and re-select from the table with the key.
Now you should find the record which in meantime was inserted by another process.

Yes, of course. I guess my question should have been, is there a standard db-independent 'unique constraint violation' exception that I can use?

Thanks,
-Gautam

I fear not.
Per definition a JDBC-Driver has to throw only exceptions of type java.sql.SQLException.
There's no common definition 'how' unique constraint violations have to be wrapped into a java.sql.SQLException.
This is vendor-specific.