For more than a year and a half I have had to use some kind of multi-node tree to do
something. However, all I found from the Internet are binary trees or equivalents.
I therefore decided to write my own. Because I had to finish it pretty fast, I spent only
two days for it, so it may be buggy.
In the project, it has three classes working like in STL container-iterator style.
template <class Key, class T> class Tree
template <class Key, class T> class Tree_iterator
template <class Key, class T> class TreeNode
A tree contains many tree nodes, while tree nodes contain data. Tree nodes of a tree are called
children. In the design each node is assigned a level and the root node is of level 1.
is for navigation within the tree. Two tree-walk styles are defined. One walks down from the root node,
while the other one use post-style to walk through all children of the tree. For each walk action,
tree_iterator is returned, so the user can use the iterator to reference the tree node and get the data.
I wrapped all the classes under the namespace Tiffany. For example, declaration of the tree with key type
wstring, and node data type of string, and add two levels of element is like:
Tiffany::Tree<wstring, string> x(L"Node 1", "Root");
px = x.AddChild(L"Key1","A");
px1 = x.AddChild(px, L"Key2", "1");
For a walk down action, you can limit it to a part of a tree
("sub-tree") by using the function
SetSubTreePivot(px), and you can set the pivot
to point back to the root node by calling SetWalkDownRootPivot. Try it out!. You can even delete part of the tree by calling
Unfortunately I didn't have much time to prepare this document, but it should
be obvious how to the classes.