When working with the metadata tree, you have two data structures to your convenience:

In the following you will find two examples of how to use the metadata tree. Suppose that the JSP templates on the CAE side have been prepared to include metadata about content. At different points throughout the JSP templates the code might look as follows:

...
<cm:metadata var="contentMetadata">
  <cm:property name="contentInfo">
    <cm:property name="title"
                 value="${self.content.title}"/>
    <cm:property name="keywords"
                 value="${self.content.keywords}"/>
  </cm:property>
</cm:metadata>

<div ${contentMetadata}>
  ...
</div>
...

In a preview document there might be multiple of such content-related metadata chunks attached to different DOM nodes. Suppose you want to gather the titles of all the contents that are included in such metadata chunks. One way to gather these titles in an array is the following:

var metadataService:IMetadataService = ... ;
var metadataTree:MetadataTree = metadataService.getMetadataTree();
var result:Array = [];
if (metadataTree.getRoot()) {
  var nodesToProcess:Array = [metadataTree.getRoot()];
}
var arrayIndex:int = 0;
while (arrayIndex < nodesToProcess.length) {
  var currentNode:MetadataTreeNode = nodesToProcess[arrayIndex];
  if (currentNode.getProperty("contentInfo")) {
    var title:String =
      currentNode.getProperty("contentInfo").title;
    result.push(title);
  }
  if (currentNode.getChildren()) {
    nodesToProcess =
      nodesToProcess.concat(currentNode.getChildren());
  }
  arrayIndex++;
}

In this example, the whole metadata tree is traversed in a breadth-first manner. For each node it has to be checked whether it has the contentInfo property as there might be metadata nodes with completely other information.

The code can be simplified considerably if a filtered metadata tree is retrieved:

var metadataService:IMetadataService = ... ;
var metadataTree:MetadataTree =
  metadataService.getMetadataTree(["contentInfo"]);
var result:Array = [];
var metadataNodesList:Array = metadataTree.getAsList();
metadataNodesList.forEach(function (node:MetadataTreeNode) {
  result.push(node.getProperty("contentInfo").title);
}

In this case, the metadata tree is filtered on retrieval, namely for metadata nodes that contain the contentInfo property. Now it is sufficient to get all metadata tree nodes as an array, walk through it and gather the content titles.