I am using the Oracle9Dialect and 3.2.0.beta9a of Hibernate Tools. I am using JPA/EJB3 and Seam.

My build will create the DDL (using hbm2ddl) for my Entity beans. I use this DDL to create the tables and constraints in Oracle (Express).

Then, if I use seam-gen generate-entities to reverse engineer a CRUD app from the database tables, I notice that my camel case isn't the same.

For example, it would seem that:

(Original Java) -> (Oracle) -> (reverse engineered Java)
replacementDate -> REPLACEMENTDATE -> replacementdate

I would think that this translation would be bidirectional, such that:

assert java_column == db2java(java2db(java_column)))


I am thinking that if the Java field name to database column name replaced a camel case transition with an underscore, then the bidirectional property would be maintained.

Is there a flag/option to get what I would expect?

Alternatively, is there a way to activate a "Preserve Case" mode where every table and column name is quoted?

Use org.hibernate.cfg.ImprovedNamingStrategy which uses _ instead of a concatenation.

With respect to quotes you could probably do that by implementing a custom reverse engineering strategy that added backticks for every colum (though I don't think that is a healthy thing ;)

Is there a way that I can specify this in the persistence.xml? (I can't seem to find a hibernate property for naming strategy.)

Never mind, found it: hibernate.ejb.naming_strategy