Implementing an AVL tree

The AVLTree class

• Two data fields

– Entry root

– Int size

The Entry class

• Five data fields

– Object element;

– char balanceFactor = ‘=‘;

• Initialized to ‘=‘

• The possible values are =, R, and L

• If the node has the same number of children, it should be =

• If the right subtree is one longer than the left it should be R

• If the left subtree is one longer than the right it should be L

– Three reference variables pointing to Entry nodes

• Left, right, parent

The add method

• We use top town design here, by letting add call other methods to do some of the work

• We use two local variables

– Entry temp = root, ancestor = null

– Temp will point to the node which is the parent of the added node

– Ancestor

• First, is the root null?

• Then, if not, should we go right or left

• When we find the place to enter the node, we put it in, then fix it in case it made the tree unbalanced.

– We call fixAfterInsertion to fix it

• Then we adjust size, and return true, if it was not a duplicate element.

The fixAfterInsertion method

• It takes two arguments, both type Entry

• Six cases

– Case 1: all ancestors of inserted have balanceFactor of '='

– Case 2: insertion in opposite subtree of ancestor's balanceFactor

– Case 3: ancestor's balanceFactor = 'R' and o > ancestor's right element

– Case 4: ancestor's balanceFactor = 'L' and o < ancestor's left element

– Case 5: ancestor's balanceFactor = 'L' and o > ancestor's left element

– Case 6: ancestor’ balanceFactor = 'R' and o < ancestor.right.element