Tree Representation in Memory: Difference between revisions

From NovaOrdis Knowledge Base
Jump to navigation Jump to search
Line 8: Line 8:
=Representing Rooted Trees as Linked Data Structures=
=Representing Rooted Trees as Linked Data Structures=


Each tree [[Tree_Concepts#Node|node]] is represented by an object that has a '''key''' attribute and '''pointers''' to other nodes, which vary depending on the tree type.
Each tree [[Tree_Concepts#Node|node]] is represented by an object that has a '''key''' attribute and '''pointers''' to other nodes, which vary depending on the tree type:


===Representing Binary Trees===
==Binary Trees==


There are three pointers:  
There are three pointers:  
Line 18: Line 18:
<center>[[Image:Tree_BinaryTree_Representation.png]]</center>
<center>[[Image:Tree_BinaryTree_Representation.png]]</center>


===Representing Trees whose Nodes have an Arbitrary Number of Children===
==Trees whose Nodes have an Arbitrary Number of Children==


One option to represent trees whose nodes have an arbitrary number of children is to extend the scheme used for representing binary tree to any class of trees in which the number of children in each node is at most some constant k: we replace the ''left'' and ''right'' attributes by ''child<sub>0</sub>'', ''child<sub>1</sub>'', ... ''child<sub>k-1</sub>''. However, this is not ideal: we cannot represent an unbounded number of children, and also when the number of children k is bounded by some large constant, but most nodes have a small number of children, we waste a lot of memory.
One option to represent trees whose nodes have an arbitrary number of children is to extend the scheme used for representing binary tree to any class of trees in which the number of children in each node is at most some constant k: we replace the ''left'' and ''right'' attributes by ''child<sub>0</sub>'', ''child<sub>1</sub>'', ... ''child<sub>k-1</sub>''. However, this is not ideal: we cannot represent an unbounded number of children, and also when the number of children k is bounded by some large constant, but most nodes have a small number of children, we waste a lot of memory.

Revision as of 21:28, 9 October 2021

Internal

Overview

Rooted trees can be implemented in several ways:

Representing Rooted Trees as Linked Data Structures

Each tree node is represented by an object that has a key attribute and pointers to other nodes, which vary depending on the tree type:

Binary Trees

There are three pointers:

  • p, which points to the parent node. Only the tree's root has a null parent.
  • left and right, which point to the left and the right child. If the node has no left child, left is NULL. If the node has no right child, right is NULL.
Tree BinaryTree Representation.png

Trees whose Nodes have an Arbitrary Number of Children

One option to represent trees whose nodes have an arbitrary number of children is to extend the scheme used for representing binary tree to any class of trees in which the number of children in each node is at most some constant k: we replace the left and right attributes by child0, child1, ... childk-1. However, this is not ideal: we cannot represent an unbounded number of children, and also when the number of children k is bounded by some large constant, but most nodes have a small number of children, we waste a lot of memory.

Another scheme to represent an arbitrary number of children is left-child, right-sibling representation. Instead of having a pointer to each of its children, each node has only three pointers:

  • parent
  • left-child points to the leftmost child
  • right-sibling points to the node's sibling immediately to the right.
Tree LeftChild RightSibling Represenation.png