The Tree Control: Managing Data with JTree

This chapter is from the book

Core JFC, 2nd Edition

Learn More Buy

This chapter is from the book

This chapter is from the book 

Core JFC, 2nd Edition

Learn More Buy

The TreePath Class

If DefaultMutableTreeNode is the physical representation of a node, then TreePath is its logical representation. Instead of pointing you directly at the node, it tells you the node's "address" within the tree. So long as the node itself and its parents are not moved within the tree, its TreePath will remain constant and can always be used to address it. TreePath has two public constructors:

public TreePath(Object singlePath) public TreePath(Object[] paths) 

A TreePath stores an array of objects that describes the path to a node. The second constructor sets up this array from its argument, while the first is a convenience method that creates an array of length 1 and initializes it with the single object passed as the argument. TreePath has the following methods:

public boolean equals(Object) 
public Object getLastPathComponent(); 
public Object getPathComponent(int index); 
public int getPathCount(); 
public Object[] getPath(); 
public TreePath getParentPath(); 
public TreePath pathByAddingChild(Object child); 
public boolean isDescendant(TreePath treePath) 
public String toString() 

None of these methods are concerned with what the objects that the TreePath stores actually are—they just treat them as anonymous objects. The equals method returns true if the object passed as its argument is another TreePath with an object array of the same length as its own and for which the result of comparing each object in its own array with the corresponding object in the other array (using the equals method the objects being compared) returns true.

The getPath method returns the array of objects installed by the constructor, while getLastPathComponent returns the last object in the array. In the case of a TreePath representing a set of directories from a file system tree ending in a file, this method would return the object that corresponds to the file itself. Similarly, getPathCount returns the number of objects in the path and getPathComponent returns a single component of that path given its index, where 0 is the component at the root. getParentPath returns a TreePath object representing the parent of the TreePath against which it is invoked. pathByAddingChild creates and returns a TreePath object that represents the results of appending the Object passed as its argument to the current TreePath. In terms of the file system tree, if you invoked this method against the TreePath for a node representing directory and passed it an Object representing a file or subdirectory within that directory, it would return a TreePath for that file or subdirectory.

The isDescendant method determines whether the TreePath passed as its argument represents a descendant of the TreePath against which it is invoked. This will be true if the TreePath passed as the argument has at least as many objects in its array as the TreePath against which this method is invoked, and each of the objects in the invoked TreePath's array is equal to the corresponding object in the array of the proposed descendant. Finally, the toString method prints a readable representation of the TreePath by invoking the toString methods of each object in its array and concatenating them, separated by commas.

The TreePath doesn't care what the objects that it manipulates are, but you need to know what they are in order to make much use of a TreePath. The objects that TreePath actually stores are the TreeNodes for each node of the tree between the root and the object that the TreePath relates to. In most cases (and in all of the examples in this chapter), this object will actually be a MutableTreeNode. This means that, in an event handler that receives a TreePath in its event, you can access the information from the data model that the event relates to, including the user object associated with the node, simply by getting a reference to the node from the TreePath. You'll see when looking at the file system tree component that is developed later in this chapter why this is a useful feature.

Let's look in more detail at what a TreePath is made up of by using an example. TreeExample2 is a development of the last example in which the tree is built and then, every 30 seconds, the TreePath object for whatever is on row 4 of the tree's display is obtained and its content is printed. Here is the interesting part of this program:

for (;;) { 
  Thread.sleep(30000); 

  // Get TreePath for row 4
  TreePath p = t.getPathForRow(4);
  if (p == null) { 
    System.out.println("Nothing on row 4!");
    continue; 
  } 

  // Print the official description of "p" 
  System.out.println("===================\n" + p); 

  // Now look inside
  Object[] o = p.getPath();
  for (int i = 0; i < o.length; i++) { 
    System.out.println("Class: " + o[i].getClass() +
       "; value: " + o[i]);
    if (o[i] instanceof DefaultMutableTreeNode) { 
     Object uo =
          ((DefaultMutableTreeNode)o[i]).getUserObject();
     if (uo != null) {
       System.out.println("\tUser object class: " +
         uo.getClass() +"; value: " + uo);
     }
    }
  }
}

The TreePath for whatever is on row 4 is obtained by using the get-PathForRow method of JTree. If the row is empty, as it will be when you start the program, this method returns null, so a message is printed and nothing further happens for another 30 seconds. Otherwise, the TreePath toString method is used to see its own description of itself, then getPath is called to get its array of component parts. For each object returned in this array, its Java class and its value are printed. Finally, if any object that is obtained from the array is a DefaultMutableTreeNode, its getUserOb-ject method is called to get the user object associated with the node and, if there is one, its class and its value are also printed. You can run this example yourself by typing:

java JFCBook.Chapter10.TreeExample2 

Following are the results of three iterations around the loop with three different directories expanded to place different objects on row 4 of the display.

 [null, Skylab, 3] 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: null
Class: class javax.swing.tree.DefaultMutableTreeNode; value: Skylab
      User object class: class java.lang.String; value: Sky-
Lab
Class: class javax.swing.tree.DefaultMutableTreeNode; value: 3 
       User object class: class java.lang.String; value: 3 
 [null, Apollo, 13]
Class: class javax.swing.tree.DefaultMutableTreeNode; value: null 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: Apollo
       User object class: class java.lang.String; value:
Apollo 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: 13
       User object class: class java.lang.String; value: 13 
 [null, Apollo, 12, Pete Conrad] 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: null 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: Apollo 
       User object class: class java.lang.String; value: 
Apollo 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: 12 
       User object class: class java.lang.String; value: 12 
Class: class javax.swing.tree.DefaultMutableTreeNode; value: Pete Conrad 
       User object class: class java.lang.String; value: Pete 
Conrad 

The toString method of TreePath prints the path to the node that was on row 4. Next, the objects that correspond to the various parts of the path are shown. As you can see, there is one object for each node in the path, starting at the root and ending with the node that corresponds to the Tree-Path. As noted above, these objects are actually the nodes that were stored in the DefaultTreeModel, so they are all instances of DefaultMutable-TreeNode and invoking toString on them prints the string that was supplied to the constructor that created them. Finally, each of these nodes, apart from the root node, has an associated user object that, as you can see, is the same string. This is not, however, a surprise, because the DefaultMutable-TreeNode constructor that was used described its argument as the user object associated with the node. In fact, the toString method of Default-MutableTreeNode just invokes the toString method of the user object if there is one and returns an empty string if there isn't. This explains why the DefaultMutableTreeNode returns the string passed to its constructor.

Incidentally, we now have the information needed to modify one of our earlier example programs so that the string root doesn't appear next to the root node on the display, as it did in Figure 10–7. In that example, the root nodes were not explicitly created by the program: Because each JTree was built by passing a Hashtable to its constructor, the root nodes were built for us. The string displayed for the root node is the result of invoking toString against that node. You now know that, because this node is a DefaultMu-tableTreeNode, its toString method will delegate to that of its user object if there is one and return an empty string otherwise. When the root nodes in Figure 10–7 were created, the string root was assigned as the user object, so to make it disappear, you just have to use the setUserObject method to set the user object to null.

Here is how you would create a tree that shows a root folder without the root label:

JTree t = new JTree(h);
t.setShowsRootHandles(false);
t.setRootVisible(true); 

// Remove the 'root' label 
Object rootNode = t.getModel().getRoot(); 
((DefaultMutableTreeNode)rootNode).setUserObject(null); 

This code gets a reference to the model from the JTree using the get-Model method and then gets the root node from the model by using the TreeModel getRoot method. This method is defined to return an Object, but you know that the root node is a DefaultMutableTreeNode, so all you have to do is cast it and use setUserObject to remove the root string. You can use the same technique to assign an arbitrary label to the root node. You can see how this works in practice using the command

java JFCBook.Chapter10.HashHandleTree3 

which produces the same result as that shown in Figure 10–7, except that the root nodes do not have any associated text.