Skip to content

sebivenlo/sebidao

Repository files navigation

sebidao

Opinionated and generic data access object implementation for postgreSQL database.

It does not use postgreSQL features per se, but it is only tested with postgreSQL.

This DAO is a work in progress. It supports the basic CRUD operations plus a few extras in a generic way, with a little help of the user-programmer.

The only required helper class to an entity say student is any mapper, like StudentMapper, which knows enough of the student class to take a student instance apart in an array of objects and do the reverse, using a constructor that is assumed (we said opinionated) an all args constructor that takes it's parameters in field declaration order. This is the same as you would get with a @Data annotated pojo with @AllArgsConstructor using lombok.

The Student class would probably be friendly and lend a helping hand to the Mapper by providing just such method and constructor, but only package private, mind you. That will do. Using a squirt of lombok juice would make that both easy and just right.

The javadoc can be found at the Fontys Venlo Javabits website under Sebi DAO javadoc

The TAO of DAO

A DAO should be simple to use. Adding complexity that serves a niche use case but increases complexity for the simple use cases is contrary to the TAO of DAO.

DAO is a pattern to reduce coupling. This implies that the application logic should not know nor care what the underlying persistence technology is. The better this objective is served, the more reuse can be achieved in the only valuable part of the software: the application logic it selves.

The current postgreSQL jdbc based implementation supports the following operations:

public interface DAO<K extends Serializable, E extends Entity2<K>> extends AutoCloseable {
    Optional<E> get( K id );
    Collection<E> getAll();
    Collection<E> getByColumnValues( Object... keyValues );
    E save( E e );
    E update( E e );
    void delete( E e );
    void delete( K k );
    TransactionToken startTransaction() throws Exception;
    DAO<K, E> setTransactionToken( TransactionToken tok );
    TransactionToken getTransactionToken();
    int size();
    int lastId();
    void close() throws Exception;
    Collection<? extends E> saveAll( Iterable<E> entities );
    Collection<? extends E> saveAll( E... entities );
    void deleteAll( Iterable<E> entities );
    void deleteAll( E... entities );
}

A minimal Mapper (since version 0.7.4) could look like this.

import java.util.function.Function;
import nl.fontys.sebivenlo.dao.AbstractMapper;

public class EmployeeMapper extends AbstractMapper<Integer, Employee> {

    public EmployeeMapper() {
        super( Integer.class, Employee.class );
    }

    @Override
    public Function<Employee, Integer> keyExtractor() {
        return Employee::getEmployeeid;
    }
}

Implementing the Object[] asParts() in the entity class (Employee in the above), can make the run time efficiency a bit better, by avoiding reflection in deconstructing the entity instance.

The Javadoc of the development branch is hosted at javabits.fontysvenlo.org