org.teiid.translator.infinispan.hotrod.metadata

Class ProtobufMetadataProcessor

java.lang.Object

org.teiid.translator.infinispan.hotrod.metadata.ProtobufMetadataProcessor

All Implemented Interfaces:

MetadataProcessor<ObjectConnection>

public class ProtobufMetadataProcessor
extends Object
implements MetadataProcessor<ObjectConnection>

The ProtobufMetadataProcess is the logic for providing metadata to the translator based on the google protobuf Descriptors and the defined class types.

Here are the rules that are being followed.

The Cache defined by the resource adapter The cache defined by the connection will be processed to create 1 or more tables. Table Each class defined in the protobuf is mapped to a table.
Attributes Get/Is methods will be become table columns.
One-to-one association The classes on both sides will be merged into a one table. The nameInSource for the attributes in the class contained within the root class will be assigned based on this format: "get{Method}" name + "." + attribute
Example: Person contains Contact (getContact) and Contact has getter method getPhoneNumber so the nameInSource for the column would become Contact.PhoneNumber
One-to-many association The classes on both sides are mapped to two tables. The root class will have primary key created. The table for the "many" side will have a foreign created.
When an attribute is included from the many side (children), the total number of rows will be multiplied by by the number of children objects.

Decisions

User will need to decide what will be the primary key. Will the key to the cache or a real attribute on the root class will be used as the primary key.