Question

I’m having trouble getting Hibernate to perform a bulk insert on MySQL.

I’m using Hibernate 3.3 and MySQL 5.1

At a high level, this is what’s happening:

@Transactional
public Set<Long> doUpdate(Project project, IRepository externalSource) {
    List<IEntity> entities = externalSource.loadEntites();
    buildEntities(entities, project);
    persistEntities(project);
}
public void persistEntities(Project project) {
     projectDAO.update(project);
}

This results in n log entries (1 for every row) as follows:

Hibernate: insert into ProjectEntity (name, parent_id,
path, project_id, state, type) values
(?, ?, ?, ?, ?, ?)

I’d like to see this get batched, so the update is more performant. It’s possible that this routine could result in tens-of-thousands of rows generated, and a db trip per row is a killer.

Why isn’t this getting batched? (It’s my understanding that batch inserts are supposed to be default where appropriate by hibernate).

Answer 1

As documented in the Chapter 13. Batch processing:

If you are undertaking batch
processing you will need to enable the
use of JDBC batching. This is
absolutely essential if you want to
achieve optimal performance. Set the
JDBC batch size to a reasonable number
(10-50, for example):

hibernate.jdbc.batch_size 20

Hibernate disables insert batching at
the JDBC level transparently if you
use an identity identifier generator.

Don’t forget to flush and then clear the session regularly or you’ll get OutOfMemoryException as documented in 13.1. Batch inserts.

But IMO, for tens-of-thousands of rows, you should consider using the StatelessSession interface.