DAO vs ORM(hibernate) pattern [closed]

Answer 1

Let me provide a source code example to hvgotcodes good answer:

public class Application
{
    private UserDao userDao;

    public Application(UserDao dao)
    {
        // Get the actual implementation
        // e.g. through dependency injection
        this.userDao = dao;
    }

    public void login()
    {
        // No matter from where
        User = userDao.findByUsername("Dummy");
    }
}


public interface UserDao
{
    User findByUsername(String name);
}

public class HibernateUserDao implements UserDao
{
    public User findByUsername(String name)
    {
        // Do some Hibernate specific stuff
        this.session.createQuery...
    }
}

public class SqlUserDao implements UserDao
{
    public User findByUsername(String name)
    {
        String query = "SELECT * FROM users WHERE name = '" + name + "'";
        // Execute SQL query and do mapping to the object
    }
}

public class LdapUserDao implements UserDao
{
    public User findByUsername(String name)
    {
        // Get this from LDAP directory
    }
}

public class NoSqlUserDao implements UserDao
{
    public User findByUsername(String name)
    {
        // Do something with e.g. couchdb
        ViewResults resultAdHoc = db.adhoc("function (doc) { if (doc.name=='" + name + "') { return doc; }}");
        // Map the result document to user
    }
}

So, as already mentioned, DAO is a design pattern to minimize coupling between your application and you backend whereas ORM deals with how to map objects into an object-relational database (which reduces coupling between the database and you application, but in the end, without using a DAO your application would be dependend on the ORM used or on a higher level a standard like JPA).

Therefore, without DAOs, it would be really hard to change your application (e.g. moving to a NoSQL database instead of a JPA compatible ORM).