WHAT IS IT:
SchemaCreator automatically creates and updates your schema. Based on meta-description of persistent entities, it:
- creates tables with their columns, PK, indexes and some constraints
- updates existing tables adding or updating columns, indexes etc.

These meta-descriptions (TableDescription) can be created by using TableDescriptionUtil.getTableDescription() 
for JPA entities or programatically for non-JPA entities.

It supports PostgreSQL and H2. Used to support Oracle too, but I stopped maintaining it a while ago.
Guess that with minor effort it would work too.

You no longer have to maintain SQL scripts per DB or manually update the schema every time you work with
a new SW version.

Thanks to https://github.com/rimar for the original idea and massive contribution!

EXAMPLE:
//your persistent entity
@Table(name = "Blog")
public class Blog {
    private static enum Status {
        GOOD, BAD
    }

    @Id
    public long id;

    @Column
    public String name;

    @Lob
    public String content;

    @Column
    public Timestamp lastUpdate;

    @Column(name = "IS_ACTIVE")
    public boolean active;

    @Column
    @Enumerated(EnumType.STRING)
    public Status enumField;
}

//put the following in your server startup sequence
    DriverManagerDataSource ds =
        new DriverManagerDataSource("jdbc:postgresql:dbmagic", "foo", "bar"); // or whatever data source you're using

    SchemaCreator schemaCreator = new SchemaCreator(ds, DBType.POSTGRESQL);

    schemaCreator.createTables(
        Arrays.asList(TableDescriptionUtil.getTableDescription(Blog.class)));
//end

Also see tests for examples

NOTICE THAT:
I don't analyze JPA relations (OneToMany, ManyToMany). Use TableDescription.addJoinTableDescription()
to declare them programatically.

If you want logging, set the org.apache.log4j.Logger programatically.

DDL's transaction support is database specific, so I let you decide whether schema creation should be transactional and
to what extent.

Tests are really scarse - I hope to write some more in the future. Moreover, running tests over PostgreSQL requires
a setup, so the tests run over H2, but there is a connection section for PostgreSQL.

N.B. You can extend schema creator and respective dialects to support more DDL stuff, like views, sequences, etc.