Refactor primary key annotations [DATACASS-361]

Mark Paluch opened DATACASS-361 and commented

Today's @PrimaryKey, @Id and @PrimaryKeyColumn lead to ambiguity which types to use.

• @Id: applied on field/property level to mark the primary key/identified
• @PrimaryKey: similar to @Id but with column name property
• @PrimaryKeyColumn: column name, ordering and primary key type properties

A typical compound key entity looks like:

@PrimaryKeyClass
class PersonKey {
 @PrimaryKeyColumn(name="first_name", type=PARTITIONED, ordinal = 1) String firstname;
 @PrimaryKeyColumn(type=CLUSTERED, ordinal = 2) String lastname;
}

@Table
class Person {
  @PrimaryKey PersonKey key;
  @Column("fieldname") String fieldname;
}

or 

@Table
class Person {
  @Id PersonKey key;
  @Column("fieldname") String fieldname;
}

and a simple key entity:

@Table
class Person {
  @PrimaryKey String key;
  @Column("fieldname") String fieldname;
}

or 

@Table
class Person {
  @Id String key;
  @Column("fieldname") String fieldname;
}

The code raises following questions:

1. When to use @Id and when @PrimaryKey?
2. Why is it that @Column cannot be applied to @Id properties?
3. PrimaryKeyColumn declares properties that can be left empty, depending on the primary key type
4. Are all annotations required?

Proposal

1. Deprecate @PrimaryKeyColumn in favor of @PartitionKey(value [optional, ordinal]) and @ClusteringKey(value [optional, ordinal], ordering) that can be combined with @Column. @PrimaryKeyColumn support can remain active and the new annotations can be built as composed annotations
2. Deprecate @PrimaryKey in favor of @Id that can be combined with @Column.

No further details from DATACASS-361