Java: how to speed up JTree reconstruction by reading data from mongodb (remote server)?

Question

basically mongodb stores each node userobject(NodePro) with parentId.

This class recursively builds a Jtree from querying all the children of a given parentId.

The problem is I have to wait up to 5 minutes for the entire tree to load when dealing with hundreds or thousands of nodes of a single tree.

Is there a way to speed this up significantly? currently even dealing with under hundred nodes, it will take a long time for the tree to become complete.

public class BuildTree {

    public BuildTree(DefaultMutableTreeNode treeNode){
        DefaultMutableTreeNode aParentNode = treeNode;
        try {
            processChildren(aParentNode);
        } catch (Exception e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
    }

    public void processChildren(DefaultMutableTreeNode parentNode) throws Exception{
        NodePro np = (NodePro) parentNode.getUserObject();      

        List<NodePro> nodelist = NodeDAO.getInstance().getChildrenOfParent(np.getId().toString());

        if (nodelist.isEmpty()){    
            System.out.println("empty");
            return;
        }else{

             for (int i=0; i< nodelist.size(); i++)
              {            

                 NodePro childnode = nodelist.get(i);
                 DefaultMutableTreeNode child = new DefaultMutableTreeNode(childnode);

                 TreeModel.getInstance().insertNodeInto(child,parentNode,TreeModel.getInstance().getChildCount(parentNode));

                 DefaultMutableTreeNode parent = (DefaultMutableTreeNode) 
                    TreeModel.getInstance().getChild(parentNode, i);

                 processChildren(parent);
              }

            return;
        }   
    }

E.g. JProfiler. Rn the project under the profiler and check where the time is wasted. — StanislavL
is there a free open source one that works with eclipse ide? — KJW

Ted M. Young Ted M. Young · Accepted Answer · 2011-08-07T02:47:06

@trashgod is right in his comment: where is the code spending its time? Is it in the object creation? DefaultMutableTreeNode doesn't perform well when you've got thousands of node to create. Is it during retrieval of the node information from MongoDB?

If you can't speed up the creation of the tree, you'll want to investigate lazily-loading the children of a node. That is, don't load the node until the user requests to see the children.

Other possibilities are to load the children nodes in a background thread.

However, like any performance improvement work, you need to measure which part of your code is slow and then figure out how to improve it.

Java: how to speed up JTree reconstruction by reading data from mongodb (remote server)?

2 Answers