Homegrown Self-Balancing Binary Search Tree (BST)

Self-Balancing Binary Search Tree (BST) in Pure VB6 Code

This is a pure VB6 self-balancing binary search tree (BST) developed in the vein of Adelson-Velsky and Landis's (AVL) work. If you're familiar with the VB6 Collection object, it has many similarities. With respect to speed, it competes well with a VB6 Collection, and offers several advantages. It's intended to be used without the need for any error trapping. The only error it should produce is possibly an out-of-memory error, but even here, each node should take less memory than a VB6 Collection node.

Also, one of the large advantages over the VB6 Collection, is that we can rather easily create nodes with any type key, and/or any type value. The initial release is the "String Key" and "no Value" category. I will develop other categories, with other types of keys and/or values, possibly upon request.

Another advantage, regardless of the category, is that we have easy access to the keys, and they'll always be sorted. So, if you like, you can use this as a sort algorithm, to which additional keys can be added, maintaining your sorting.

And, this one is full Unicode case-sensitive. The fact that the Collection is case-insensitive has been a source of consternation for me on many occasions.

To use this "String Key" and "no Value" category, drop both the BST_StrKey_NoVal.cls and Nod_StrKey_NoVal.cls into your project, create an object variable of BST_StrKey_NoVal type, instantiate it, and you're up and running. Don't do anything with the Nod_StrKey_NoVal class, as it's only used internally by the BST_StrKey_NoVal class.

You can instantiate as many different BST_StrKey_NoVal objects (i.e., trees) as you need. They will be maintained as separate BST trees.

The following are the exposed procedures available for production code. (For illustration purposes, I'll assume your instantiated object variable is named "BST".):

• BST.Insert(Key)
  
  Key must be a non-empty string. No duplicates allowed. Returns TRUE if successfully added, or FALSE if it's a DUPE.
• BST.Delete(Key)
  
  Key must exist in the tree. Returns TRUE if deleted, or FALSE if not found.
• BST.KeyExists(Key)
  
  Returns TRUE or FALSE.
• BST.Count
  
  Returns the number of nodes currently in the tree.
• BST.Clear
  
  Clears out all the nodes, without any need for re-instantiating the BST object.
• BST.KeyArray
  
  This returns a sorted string array of all the nodes in the tree. There is an optional "descending" argument.
• BST.NextKey
  
  This provides a way to loop through the keys in the tree (sorted).
• BST.PrevKey
  
  Same as BST.NextKey, but in descending order.
• BST.Reset
  
  If you're actively using either BST.NextKey or BST.PrevKey, this provides a way to start the internal loop over.

A few more notes about these procedures:

• The way BST.NextKey and BST.PrevKey are designed, they can be used in conjunction with each other. For example, if we call BST.NextKey: BST.PrevKey: BST.NextKey, those two NextKey calls will return the same key.

• Anytime BST.Insert or BST.Delete are called, the BST.NextKey and/or BST.PrevKey will be reset.

• If you'd like to use the For Each... syntax, it can be done with the KeyArray property:

That's about it, for now. There's a project attached with a substantial front-end for illustration and testing. However, again, all you need to use this is the two CLS modules. Enjoy. :)

Change Log:

Attached Files

• BST_StrKey_NoVal.zip (12.9 KB)