de.hybris.platform.webservices.util.objectgraphtransformer

Annotation Type GraphNode

@Retention(value=RUNTIME)
public @interface GraphNode

Required Element Summary

Required Elements

Modifier and Type	Required Element and Description
java.lang.Class	target Maps this source class to a target class.

Optional Element Summary

Optional Elements

Modifier and Type	Optional Element and Description
java.lang.Class[]	addNodes Manually adds additional nodes.
java.lang.Class<? extends NodeFactory>	factory A finder can be used as replacement for a target class.
java.lang.String	uidProperties The runtime values of these properties (as comma separated list) are taken to calculate a unique id for that node type.

Element Detail

target

public abstract java.lang.Class target

Maps this source class to a target class. Target class must have a default constructor. A target instance is created whenever a source to target transformation is requested but no target parameter is available.

Example: AbstractGraphTransformer.transform(Object)

Returns:

target class

factory

public abstract java.lang.Class<? extends NodeFactory> factory

A finder can be used as replacement for a target class. A finder must implement NodeFactory. Finders are allowed to return 'null'. In that class configured target class is taken.

Returns:

NodeFactory class

Default:

de.hybris.platform.webservices.util.objectgraphtransformer.NodeFactory.class

addNodes

public abstract java.lang.Class[] addNodes

Manually adds additional nodes. Generally nodes are detected automatically but in case nodes are elements of collection elements they can't be recognized.

Returns:

list of additional nodes

Default:

void.class

uidProperties

public abstract java.lang.String uidProperties

The runtime values of these properties (as comma separated list) are taken to calculate a unique id for that node type. When no such properties are configured, 'uid' would be the objects hashcode. If properties are defined, 'uid' would be created by combining the ID of each property value.

To understand the purpose of why these may be useful some basics must be known:
Graph processing recognizes infinite graphs by just looking, whether a graph node was already processed. This is done by maintaining a lookup-map of already processed node instances. Graphs with infinite loops/cycles are therefore no problem.
This is basically a very easy solution and depends on node identity.

There might be another request, which needs nodes to be recognized as "already processed" based on equality. For instance when using external frameworks which building up a graph (like JAXB) a graph gets created always with new instances, no node-instance is reused twice although it might logical be the case.
In other words: XML structures can form infinite graphs too, but JAXB can't handle that and creates always new node instances instead of recognizing and sharing already created nodes.

This is what can be done here, by just defining a set of properties whose values are taken for equality check.

Returns:

uid

Default:

""