My B+Tree can compile but have runtime error ... i can't tink why ...

Question

Sorry to bother anyone ... i have spent almost 3 weeks trying to debug the problem...
My B+ Tree can compile but can't run as it throw invalid Null pointer Exception ..
I suspected the error is in the split (int order ) method in the BTreeNode class but i can't remedy for it
This my program code
public class BTree {
* The order of the B-Tree. All nodes of the B-Tree can store up to 2*order key values.
private int order;
* The number of key (with associated data values) stored in the B-Tree.
private int count;
* The root node of the B-Tree
private BTreeNode root;
* The first leaf node in the B-Tree. This field is used to give the starting point for sequential
* access to the keys and data stored in the leaf nodes.
private BTreeNode first;
* The last leaf node in the B-Tree. This field is used to give the starting point for reverse
* sequential access to the keys and data stored in the leaf nodes.
private BTreeNode last;
* A change count that can be used to invalidate iterators. This field is updated whenever a key plus data pair
* is added to the B-Tree (or the data associated with a key is changed), or when a key plus data pair are
* deleted from the B-Tree.
private int changeCount = 0;
// WHEN DOING ASSIGNMENT 5, DO NOT ADD ANY ADDITIONAL FIELDS TO THIS CLASS
// You will loose marks if you add additional fields to this class. The fields above are all that
// you need. If you need a variable in a method, use a local variable. I have seen too many
// assignments where students add fields rather than create local variables. Hopefull the threat
// of loosing (quite a few) marks will help reduce this habit.
* A general exception class for errors when constructing or manipulating a B-Tree. Use the string
* parameter to the constructor to say what the error really is.
public class BTreeError extends RuntimeException {
public BTreeError(String reason) {
super(reason);
* A constructor that creates an empty B-Tree of the given order.
* <p/>This is the only constructor provided at the moment for this BTree class. Could consider
* adding the equivalent of a 'copy constructor' that creates a new BTree object from an existing
* BTree object.Constructor
* creates the root of a btree
* A constructor that creates an empty B-Tree of the given order.
* <p/>This constructor need to copy the order parameter to the field of same name, and initialise the
* root, cound, first and last fields of the BTree object.
* @param order The order of the BTree.
public BTree(int order) {
count = 0;
this.order = order;
root = new BTreeNode(true, null, -1, null, null);
first = root;
last = root;
* A method to return a SequentialIterator object that is positioned just before the first key
* of the BTree object.
* <p/>Do not modify this method.
* @return A SequentialIterator object.
public SequentialIterator iterator() {
return new BTreeIterator();
* A mehtod to return a SequentialIterator object that is positioned at a key found through a call
* to the searchFor() method.
* <p/>Do not modify this method.
* @param position A SearchPosition() object that usually has been returne by a call to the searchFor() method.
* @return A SequentialIterator object initialised to start at particular key in the BTree.
public SequentialIterator iterator(SearchPosition position) {
return new BTreeIterator(position);
* A method to return a string representationo the BTree.
* <p>The format of the string is:
* <pre>BTree: Order=<digits>, size=<digits>, root=<BTreeNode string></pre>
* <p/>Do not modify this method.
* @return A string to represent the BTree
public String toString() {
StringBuffer s = new StringBuffer("BTree: Order=");
s.append(order).append(", size=").append(size()).append(", root=").append(root.toString());
return s.toString();
* Method to determine the number of records stored in the B-Treee.
* <p/>Do not modify this method
* @return the number of records in the B-Tree.
public int size() {
return count;
* method to return the order of the B-Tree.
* <p/>
* <p>This is the smallest number of key values for all nodes
* except the root node. The maximum number of key values in a node is 2*order, and the maximum number
* of child nodes for a node is 2*order+1.
* <p/>Do not modify this method.
* @return The order of the B-tree.
public int order() {
return order;
* Insert a key plus data value into the BTree.
* <p/>This method needs to locate the leaf node in which the key + data value should be
* inserted, and then call the insertLeaf() method of BTreeNode to do the insertion.
* <p/>This method will always result in a change to the BTree, so it should increment
* the change count.
* <p/>The method may result in only the data associated with an existing ke being changed,
* so incrementing the count field should be done in the BTreeNode method (if needed).
* <p/>This is one of the method you need to complete for assignment 5.
* @param key The key associated with the data value to be added to the B-Tree
* @param data The data value to be added (with it's associated key) to the B-Tree.
public void add(Comparable key, Object data) {
// you need to add the code for this method
// i added
BTreeNode btNode = root;
while (!btNode.isLeaf) {
int i=0;
while(key.compareTo(btNode.keys) > 0) {
i++;
if (i == btNode.numberOfKeys) break;
btNode = btNode.childNodes[i];
btNode.insert(key,data);
if (root.numberOfKeys == order*2-1) root.split(order);
* insert a object with the given key into the tree
//KeyNode keyNode = new KeyNode(key, data);
// BTreeNode keyNode = new BTreeNode(key,data);
BTreeNode btNode = root;
while (!btNode.isLeaf) {
int i=0;
while(key.compareTo(btNode.key(i)) > 0) {
i++;
if (i == btNode.numberOfKeys())
break;
btNode = btNode.child(i); }
System.out.println("hmm1");
btNode.insert(key,data );
System.out.println("hmm2");
if (root.numberOfKeys == order*2-1)
System.out.println("hmm3");
root.split(order);
System.out.println("hmm4");
* This method searches the B-Tree for an occurence of the key in a leaf node and returns the result
* of the search in a SearchPosition object.
* <p/>Note that the key may occur in an interior node of the BTree without occuring in the leaf
* nodes. This can be the result of a deletion operation. This method need to search down to the
* leaf node that should contain the key if the key and associated data is in the B-Tree, and then
* scan through the keys in the leaf node for the search key.
* <p/>The result of the search is returned as a SearchPosition object, as this allow the return
* of the success or failure of the search, as well as the data belonging to the key. It also
* allows position information to be returned so that an interator can be initialised with the
* key as the starting position for subsequent sequential access to adjacent keys.
* <p/>This is one of the method you need to implement.
* <p/>Implementation nodes:<br>
* You need to find the leaf node that may contain the key (same code as for add()), then
* scan the leaf BTreeNode for the search tree. You can do this within this method, as you
* have total access to the fields and methods of BTreeNode (as BTreeNode is an inner class
* of BTree). If you find the key, construct and return a SearchPosition object with the appropriate
* details of the position, otherwise construct add return a SearchPosition object that indicates the
* search failed.
* @param key The key to search for in the BTree
* @return A SearchPosition object returning the data and position information for the search
public SearchPosition searchFor(Comparable key) {
// You need to add the code for this method. The code below simply creates a
// SearchPosition object which indicates an unsuccessful search.
return new SearchPosition(false, null, -1);
* A method to delete a node from the BTree.
* <p/>The method should return the data object that was deleted when the key plus data object pair
* are deleted from the B-tree.
* <p/>The method should throw a BTreeError exception (with an appropriate reason string) if the
* key is not found in the B-tree.
* <p/>This is a method you can implement for bonus marks in Assignment 5.
* <p/>Implementation notes:<br>
* The easiest way to proceed is to use searchFor() to determine if they key is in the BTree, and
* (if it is in the B-tree) to return position information about the key. Throw an exception if the
* key is not in the B-tree, otherwise keep a copy of the data assocaited with the key (to return),
* then for the leaf node containing the key (a BTreeNode object), call the deleteLeafNodeKey() method,
* passing across the leaf key index of the key (so you don't have to find it again in the leaf node).
* After this method deletes the key, return the data you saved as the method result.
* @param key The key to delete (along with it's associated data) from the B-tree.
* @return The data associated with the key that was deleted.
public Object delete(Comparable key){
// You need to add the code for this method.
return null;
* The inner class BTreeNode is used to represent the nodes in the B-Tree.
* <p/>The nodes in the BTree are of two types:
* <ol>
* <li>Leaf nodes that contain the keys and associated data values, stored in ascending key order.<br>
* These leaf nodes have next and previous pointers to adjacent leaf nodes to allow an easy
* implementation of an iterator class to provide bi-directional sequential access to the keys stored
* in the BTree nodes.
* <li>Interior nodes that contain keys and links to child nodes (that are either all internal nodes
* or all leaf nodes), organised as the node of a multi-way search tree. The interior nodes have
* one more child node link than keys. The child node link at index k is to a node with keys that
* are all less than the key at index k in this node. The link at index k+1 is to a child node
* with keys that are all greater than or equal to the key at index k.
* </ol>
* The BTreeNode class allows you to create these two 'types' of nodes, depending on the parameters
* passed to the constructor.
* <p/>There are methods that should only be called for leaf nodes and methods that should only be
* called for interior nodes. These methods should throw an exception if called by the wrong node
* type. This class should really be designed using inheritance to mimic the pascal/C++ variant
* record structure, but this design is relatively easy to understand and to implement.
* <p/>Note that this class is an inner class of BTree, and so all objects will have an implict
* reference to the BTree container object. This class has direct access to all the fields of the
* BTree contaner object. In particular, the order of the BTree is available, hence this class
* does not need to keep a copy of the order as a field.
* <p/>Skeleton class provided for Objects and Algorithms Assignment 5
* <p/>Only modify the methods where the JavaDoc indicates that you need to provide code.
* @author a.sobey
* @version 1.0
* Date: 16/05/2005
public class BTreeNode {
* The actual number of key values stored in the BTreeNode. <br>Note that the BTree node has an implicit
* reference to the containing BTree object, and the maximum number of nodes that can be stored in a
* a BTreeNode (except temporarily during the split operation) is twice the <i>order</i> of the BTree.<br>
* This field is valid for both internal and leaf nodes.
private int numberOfKeys = 0;
* The array of pointers to child nodes of this node. Only <i>(numberOfKeys+1)</i> are valid if <i>numberOfKeys</i>
* is non-zero.<br>
* This array is only valid and created for internal nodes - this array is not created for leaf nodes.<br>
* There is space in the array created for one additional child node link - this makes the coding for
* splitting of an internal node easier to implement.
private BTreeNode[] childNodes;
* A reference to the parent node of this node.<br>
* This link is null if this node is the root node of the tree of BTreeNodes.<br>
* This node is valid for both internal and leaf nodes.
private BTreeNode parent;
* The index in the parent node's array of links (the <i>childNodes</i> array) for the link to this node.<br>
* This value should be set to -1 if this node is the root node (and so has no parent node).<br>
* This field is valid for both internal and leaf nodes.
private int parentIndex;
* A link to the next leaf node in the B-tree, provided to allow easy sequential access of the keys
* and values stored in the B-tree.<br>
* This field is only valid if the node is a leaf node. For non-leaf nodes set the value to null.<br>
* For leaf nodes, set the value to null if this node is the last leaf node in the B-tree.
private BTreeNode next;
* The link to the previous leaf node in the B-tree, provided ot allow easy reverse sequential access of the keys
* and values stored in the B-Tree.<br>
* This values should be set to null if this node is a leaf node but is the first leaf node in the B-Tree, or
* if this node is not a leaf node.<br>
* This field is only used in leaf nodes.
private BTreeNode previous;
* An array of comparable key objects that are stored in this node of the B-tree.<br>
* Only the first <i>numberOfKey</i> values in the array are valid.<br>
* The maximum number of keys in a node is 2*<i>order</i>, however there is space in this array
* for one additional value to make the coding of the node splitting operation easier to implement.<br>
* This field is valid for both internal and leaf nodes.
private Comparable[] keys;
* An array of data values associated with the keys stored in this leaf node of the B-tree.<br>
* Only the first <i>numberOfKey</i> values are valid.<br>
* The maximum number of data values in a node is 2*<i>order</i>, however there is space in this array
* for one additional value to make the codingof the leaf node splitting operation easier to implement.<br>
* This field is only valid for leaf nodes - for interior nodes this array is not created.
private Object[] data;
* A boolean value to indicate if the node is a leaf node or not. The structure of the remainder of the node
* depends on this value (would be nice to have variant records in Java ...).<br>
* This field is valid for both leaf and internal nodes.
private boolean isLeaf;
private int order;
* The constructor for a BTreeNode.
* <p/>The code for the constructor is provided - do not modify this constructor.
* @param isLeaf True if this node is a leaf node.
* @param parent A link to the parent node or null if this node is the root node of the B-Tree
* @param parentIndex The index of the link in the array of child node linkes in the parent node that points to this node.
* @param previous A link to the previous leaf node for sequentail access, or null if not a leaf node or no previous leaf nodes.
* @param next A link to the next leaf node for sequential access, or null if not a leaf node or the last leaf node.
public BTreeNode(boolean isLeaf, BTreeNode parent, int parentIndex, BTreeNode previous, BTreeNode next) {
this.parent = parent;
this.parentIndex = parentIndex;
this.previous = previous;
this.next = next;
this.isLeaf = isLeaf;
if (isLeaf)
data = new Object[2 * order + 1];
else
childNodes = new BTreeNode[2 * order + 2];
keys = new Comparable[2 * order + 1];
public BTreeNode( int order, BTreeNode parent)
this.order = order;
this.parent=parent;
this.keys = new Comparable[2*order-1];
this.data = new Object[2*order-1];
this.childNodes=new BTreeNode[2*order];
this.isLeaf=true;
* Returns the number of keys in this BTreeNode. Note that within the code in BTree you have access
* to all the fields of BTreeNode, so this method is not strictly necessary.
* @return The number of keys in this BTreeNode object.
public int numberOfKeys() {
return numberOfKeys;
* Returns the container BTree object for this BTreeNode object. You may like to check that container objects
* are the same when manipulating two BTreeNode objects.
* @return the containing BTree object.
public BTree container() {
return BTree.this;
* A private method to return a string representation of the array <i>keys</i>. This method is used in
* the toString() method for this class.<br>
* Do not modify the code provided for this method.
* @return A string representation of this nodes array of keys.
private String keyString() {
StringBuffer s = new StringBuffer("{");
for (int index = 0; index < numberOfKeys; index++)
s.append(index > 0 ? "," + keys[index] : keys[index]);
return s.append("}").toString();
* A private method to return a string representation of the array of data values stored in a leaf node.<br>
* This method is used in the toString() method of BTreeNode. The method does not check if this node is a
* leaf node, as it is not intended to be called directly from outside of this class, and the toString()
* method only calls this method if the node is a leaf node.<br>
* Do not modify the provided code for this method.
* @return a string representation of the data values array of a BTreeNode.
private String dataString() {
StringBuffer s = new StringBuffer("(");
for (int index = 0; index < numberOfKeys; index++)
s.append(index > 0 ? "," + data[index] : data[index]);
return s.append(")").toString();
* A private method to return a string prepresentation of the array of child node links in an interior node.<br>
* This method is used in the toString() method. This method does not check if this node is an interior
* node, so you must take care to only call this method for interior nodes.<br>
* Do not modify the provided code for this method.
* @return A string representation of the array of child nodes of this BTreeNode.
private String childString() {
StringBuffer s = new StringBuffer("<");
for (int index = 0; index < numberOfKeys + 1; index++)
s.append(childNodes[index] + (index == numberOfKeys ? "" : ","));
return s.append(">").toString();
* The toString method provides a string representation of a BTreeNode.<br> This particular method does not
* include the details of all the fields of a BTreeNode. While debugging your code, you may like to include
* more information (such as the parentIndex value), but in your final submission you must have the code
* as provided in the skeleton BTreeNode class provided to you.
* @return A string representation of a BTreeNode.
public String toString() {
if (isLeaf)
return (new StringBuffer("[")).append(numberOfKeys)
// .append(',').append(parentIndex) // uncomment this line if need to check parentIndex values
.append(',').append(keyString()).append(',').append(dataString()).append(']').toString();
else
return (new StringBuffer("[")).append(numberOfKeys)
//.append(',').append(parentIndex) // uncomment this line if need to check parentIndex values
.append(',').append(keyString()).append(',').append(childString()).append(']').toString();
* Returns the key with the given index in this node. Throws a BTreeError exception if the index is not valid.<br>
* Do not modify this provided code.
* @param index The index of the key.
* @return The key value at the given index.
public Comparable key(int index) {
if (index < 0 || index >= numberOfKeys)
throw new BTreeError("Key index out of range - value = " + index);
return keys[index];
* Returns the child node at the provided index into the childNodes array.<br>
* A BTreeError exception is thrown if the node is not an internal
* node or if the index is not valid.
* <p/>Note that the child node returned will have keys that are all less than the key stored
* in the <i>keys</i> array of this node at the given index value (except the last childNode
* at index numberOfkeys, as this node has keys that are all greater than or equal to the last
* key value stored in this node).<br>
* Do not modify the provided code for this method.
* @param index The index into the array of child nodes for this internal BTreeNode object.
* @return The child node link.
public BTreeNode child(int index) {
if (isLeaf) throw new BTreeError("child() called for a leaf node");
if (index < 0 || index > numberOfKeys)
throw new BTreeError("Child node index out of range - value = " + index);
return childNodes[index];
* Returns the data value associated with the key at the given index. An BTreeError exception is thrown if the
* node is not a leaf node or if the index is invalid.
* <p/>Do not modify the provided code for this method.
* @param index The index of the key assocaited with the data value.
* @return The data value associated with the key with given index.
public Object data(int index) {
if (!isLeaf) throw new BTreeError("data() called for an internal node");
if (index < 0 || index >= numberOfKeys)
throw new BTreeError("Data index out of range - value = " + index);
return data[index];
* This method is used to determine if this node is a leaf node.
* @return True if this node is a leaf node.
public boolean isLeaf() {
return isLeaf;
* Inserts the (key, data) pair into this BTreeNode object.
* <p/>You must supply the code for this method.
* <p/>Implementation notes:<br>
* <ol>
* <li>Throw an exception if this node is not a leaf node.
* <li>Scan the keys array for index of the key greater than or equal to the insertion key.
* <li>If the key at the index is equal to the insertion key, update the data field and return - you are done.
* <li>Otherwise shuffle the keys values from the insertion index along to make a hole for the insertion key,
* and insert the insertion key into the keys array. Do the same for the data array values to insert the
* new data value into the data array at the insertion index.
* <li>increment the number of keys, and increment the container BTree object's count field.
* <li>If the number of keys in the node is now no more than 2*order, you are done, so simply return.
* <li>Otherwise the node has (2*order+1) key values, and need to split. The split operation leaves the first
* <i>order</i> keys and data values in this node (and so the node's numberOfKeys value will become
* <i>order</i>, and moves the remaining (order + 1) keys and data values to a new BTreeNode leaf node
* that you need to create.<br>
* You need to fix up the previous and next fields of this leaf node and the new leaf node you have created.<br>
* Two sub-cases:
* <ol>
* <li>If this node is the root node (i.e., it does not have a parent node), the split of this node will create
* a new root node, with a single key (the key at index (order+1)) and this node and the new BTreeNode as
* the child nodes. In my solution I used a call to the method newRootNode to do this. The newRootNode()
* method will also be used when a split of an interior node creates a new root node. See the JavaDoc for
* details of what the newRootNode() method should do. Once the new root node has been created, and all
* the fields updated due to the split, you are done.
* <li>Otherwise we need to insert in this node's parent node the middle key (at index (order+1) and the
* new node that we created. This is done by the method insertInterior(). The method is passed the
* key to insert (at location this.parentIndex in the keys array of the parent node), the index to
* to insert the key (this.parentIndex), and the new leaf node (that will be inserted at index
* (this.parentIndex+1) in the parent node's child links array).
* </ol>
* </ol>
* @param key The key to insert into the leaf node.
* @param data The key's corresponding data value.
public void insertLeaf(Comparable key, Object data) {
// You need to provide the code for this method.
// BTreeNode temp = new
int size = this.data.length;
int counter = 0;
this.keys[size] = key;
this.data[size] = data;
sort(size);
public int compareTo(Comparable o2) {
// Integer o1 = (Integer) o2;
return (((Integer)o2).compareTo(this));
*split()
*Splits a node into to nodes. This can only be done, if the node is full
*The midlest key go up into the parent, the left ones of them rest in
*this node, and the right ones go into a new node.
private BTreeNode split(int order) {
if (numberOfKeys == order*2-1) {
BTreeNode right = null;
if (parent == null) { // algo for the root-node
BTreeNode left = new BTreeNode(order, this);
right = new BTreeNode(order, this);
for (int i=0; i<order-1; i++) {
left.keys[i] = keys[i];
left.data[i] = data[i];
right.keys[i] = keys[order+i];
right.data[i] = data[order+i];
if (!isLeaf()) {
for (int i=0; i<order; i++) {
left.childNodes[i] = childNodes[i];
left.childNodes[i].parent = left;
right.childNodes[i] = childNodes[order+i];
right.childNodes[i].parent = right;
left.isLeaf = false;
right.isLeaf = false;
} else isLeaf = false;
keys[0] = keys[order-1];
numberOfKeys = 1;
left.numberOfKeys = order-1;
right.numberOfKeys = order-1;
for (int i=1; i<order*2-1; i++) {
keys[i] = null;
data[i] = null;
childNodes[i+1] = null;
childNodes[0] = left;
childNodes[1] = right;

Answer 1

* Don't post that much code. There should never be a reason to. You should be able to break your code down into small enough pieces that you can post a small example that demonstrates your problem.
* When you do post code, use [code] and [/code] tags to make it readable. You can use the code button on the message entry page.
* The stack trace will tell you which line the NPE occurred on, where it was called from, where that was called from, etc. You can use that to help you find your error. Look at the line it's complaining about. What references are on that line followed by a dot, and what arrays do you try to access the elements of. One of those must be null.
* Now that you know what[b] is null, put in a bunch of print statements to track [b]how it got to be null.