如何存儲tree structure
兩種方式
The Adjacency List Model
Modified Preorder Tree Traversal
The Adjacency List Model
存儲親子關係，透過recursive方式印出 tree
缺點為缺乏效率，遇到large tree要作多次的recursive call。
Modified Preorder Tree Traversal
每個node多存儲左、右編號，編號順序是依照preorder (中、左、右順序走訪)
Table內記錄，可以沒有parent column
缺點: 較難懂
詳細說明參考 Storing Hierarchical Data in a Database
Ternary Search Trees結合了binary search trees 及 digital search tries的優點，也就是說具備digital search tries只儲存word list節省空間與binary search tree搜尋快速的特性。適合拿來當作查詢字典的資料結構。實作在此架構的演算法除了exactly match之外，還有Partial-Match Searching & Near-Neighbor Searching。
參考資料:
Ternary Search Trees – Dr. Dobb’s Journal April 1998
Ternary Search Trees -Jon Bentley and Robert Sedgewick
Java : Plant your data in a ternary search tree
Sourceforge -C++ Code Ternary Search Trees