Disagreement is fine by me :). However, I think you really missed my point. I am not having a go at the Hibernate team for the current state of SchemaUpdate. I understand I am using it in ways that were not intended, and have to expect some problems. I even say in the post that I don’t expect them to jump to implement this feature. My problem is that they should not be telling me and other users we don’t need to do things which we certainly do need to do (like automatically update a production schema).

I also do not understand your point in your discussion about the persistence and database layer. Hibernate allows me to add index information to my mapping files, in which case the indices are created when using Hibernate to generate a fresh schema (using SchemaExport). Maybe there another purity argument against this but the fact is that it works well, reduces duplication of information in our app and allows us more easily to support multiple databases. It is just a shame that even though the information is there and understood by SchemaExport that SchemaUpdate doesn’t also use it.

In your case of needing an index… that has nothing to do with the persistence layer… the layer cannot utilize that information for anything as a query cannot be pre-planned by a client. The index was a performance improvement in the database layer only and was thought out by you in a DBA role.

I’m sure that your schema update functionality takes the old schema into account where a framework like Hibernate cannot as it only knows about the current schema.