haihongyuan.com
海量文库 文档专家
全站搜索:
您现在的位置:首页 > 幼儿教育 > 幼儿读物幼儿读物

e11

发布时间:2014-01-01 11:01:21  

Chapter 11 MULTIWAY TREES
1. Orchards, Trees, and Binary Trees 2. Lexicographic Search Trees: Tries 3. External Searching: B-Trees 4. Red-Black Trees

Pointers and Pitfalls

11.1 On the Classification of Species
1. On the Classification of Species
Definition:

?A (free) tree is any set of points (called vertices) and any set of pairs of distinct vertices (called edges or branches) such that (1) there is a sequence of edges (a path) from any vertex to any other, and (2) there are no circuits, that is, no paths starting from a vertex and returning to the same vertex.

2. Ordered Trees
?A rooted tree is a tree in which one vertex, called the root, is distinguished. ?An ordered tree is a rooted tree in which the children of each vertex are assigned an order.

?A forest is a set of trees. We usually assume that all trees in a forest are rooted.
?An orchard (also called an ordered forest) is an ordered set of ordered trees.

See pg.522 Fig.11.1

3. Forests and Orchards ?Multiple links ?first child and next sibling links ?Correspondence with binary trees

Recursive Definitions
Definition A rooted tree consists of a single vertex v, called the root of the tree, together with a forest F , whose trees are called the subtrees of the root. A forest F is a (possibly empty) set of rooted trees. Definition An ordered tree T consists of a single vertex v, called the root of the tree, together with an orchard O,whose trees are called the subtrees of the root v. We may denote the ordered tree with the ordered pair T = {v,O}. An orchard O is either the empty set ;, or consists of an ordered tree T , called the first tree of the orchard, together with another orchard O’ (which contains the remaining trees of the orchard). We may denote the orchard with the ordered pair O = (T ,O’).

4. The Formal Correspondence
Definition A binary tree B is either the empty set ; or consists of a root vertex v with two binary trees B1 数学归纳法 and B2 . We may denote the binary tree with the ordered triple B = [v;B1;B2]. Theorem 11.1 Let S be any finite set of vertices. There is a one-to-one correspondence f from the set of orchards whose set of vertices is S to the set of binary trees whose set of vertices is S.

Proof. Define f (?) = ? Define f ({v,O1},O2 ) = [ v,f (O1) ,f (O2) ] Show by mathematical induction on the number of vertices that f is a one-to-one correspondence.

5.Rotations
? Draw the orchard so that the first child of each vertex is immediately below the vertex. ? Draw a vertical link from each vertex to its first child, and draw a horizontal link from each vertex to its next sibling. ? Remove the remaining original links. ? Rotate the diagram 45 degrees clockwise, so that the vertical links appear as left links and the horizontal links as right links.

6. Summary
Orchards and binary trees correspond by any of: ? first child and next sibling links, ? rotations of diagrams, ? formal notational equivalence.

11.2 Le

xicographic Search Trees: Tries
1. Tries Definitions Definition: A trie of order m is either empty or consists of an ordered sequence of exactly m tries of order m. pg.530

2. Seach for a Key

pg.530

3. C++ Tries Declarations

pg.531-532

?Every Record has a Key that is an alphanumeric string. ?Method char key letter(int position) returns the character in the given position of the key or returns a blank, if the key has length less than position.

?Auxiliary function int alphabetic order(char symbol) returns the alphabetic position of the character symbol, or 27 for nonblank, nonalphabetic characters, or 0 for blank characters.

const int num_chars = 28; struct Trie_node { Record *data; Trie_node *branch[num_chars]; Trie_node(); };

data members constructors

class Trie { public: Trie(); bool empty(); Error_code trie_search(const Key &target, Record &x) const; Error_code insert(const Record &new_entry);

void inorder(void (*visit)(Record *)); int size(); int height(); void clear(); ~Trie(); void depth_traverse(void (*visit)(Record *)); Error_code remove(const Key &target); private: Trie_node *root; };

Move down for a blank Move to Terminate search the appropriate branch indexes letters the &target, in next character for a NULL location Error_code of key Trie::trie_search(const Key the target Record &x) of the trie of the target

4. Searching a Tries

pg.532 search Terminate

/* Post: If the search is successful, a code of success is returned, and the output parameter x is set as a copy of the Trie's record that holds target. Otherwise, a code of not_present is returned. Uses: Methods of class Key. */ { int position = 0; char next_char; Trie_node *location = root; while (location != NULL && (next_char=target.key_letter(position)) != ' ') { location=location->branch[alphabetic_order(next_char)]; position++; }

if (location != NULL && location->data != NULL) { Create a new x = *(location->data); return success; } empty Trie else return not_present; }

5. Insertion into a Tries

pg.533

Error_code Trie::insert(const Record &new_entry) /* Post: If the Key of new_entry is already in the Trie, a code of duplicate_error is returned. Otherwise, a code of success is returned and the Record new_entry is inserted into the Trie. Uses: Methods of classes Record and Trie_node.*/ { Error_code result = success; if (root==NULL) root = new Trie_node;

int position = 0; char next_char; At this point, we have tested Trie_node *location = root; for all non blank while (location != NULL && characters of (next_char=target.key_letter(position)) != ' ') new_entry. { int next_position = alphabetic_order(next_char); if (location->branch[next_position] == NULL) location->branch[next_position]=new Trie_node; location = location->branch[next_position]; position++; } if (location->data != NULL) result = duplicate_error; else location->data = new Record(new_entry); return result;

}

indexes letters moves through of new_entry the Trie

The number of steps required

to search a trie or insert into it is proportional to the number of characters making up a key, not to a logarithm of the number of keys as in other tree-based searches.

6. Deletion from a Tries

pg.533

Error_code Trie::trie_search(const Key &target, Record &x) /* Post: If the search is successful, a code of success is returned, and the output parameter x is set as a copy of the Trie's record that holds target. Otherwise, a code of not_present is returned. Uses: Methods of class Key. */ { int position = 0; char next_char; Trie_node *location = root; while (location != NULL && (next_char=target.key_letter(position)) != ' ') { location=location->branch[alphabetic_order(next_char)]; position++; }

11.3 External Searching: B-Trees
1. Multiway Search Trees



? An m-way search tree is a tree in which, for some integer m called the order of the tree, each node has at most m children. ? If k≤m is the number of children, then the node contains exactly k-1 keys, which partition all the keys into k subsets consisting of all the keys less than the first key in the node, all the keys between a pair of keys in the node, and all keys greater than the largest key in the node.

定义:一颗m阶查找树,或者是一颗空树,或者是满 足如下性质的m路查找树: 1)每个结点最多有m棵子树、m-1个关键字,其结构 为:
其中n为关键字个数,Pi 为指向子树根结点的指针, Ki为关键字。 2) Ki<Ki+1, 3) 子树Pi中的所有关键字均大于Ki、小于Ki+1, 4) 子树P0中的关键字均小于Ki,而子树Pn中的所 有关键字均大于Kn。 5)子树Pi也是m阶查找树,0≤i≤n。

2. Balanced Multiway Trees (B-Trees) Definition A B-tree of order m is an m-way search tree in which 1. All leaves are on the same level. 2. All internal nodes except the root have at most m nonempty children, and at least ?m/2 ? nonempty children. 3. The number of keys in each internal node is one less than the number of its nonempty children, and these keys partition the keys in the children in the fashion of a search tree. 4. The root has at most m children, but may have as few as 2 if it is not a leaf, or none if the tree consists of the root alone.

定义: 一棵m阶的B-树,或为空树,或为满足下列特性的m
叉树: ① 每个结点至多有m棵子树; ② 除根结点和叶结点外,其它每个结点至少有?m/2?棵子树; ③ 除非根结点为最低层非终端结点,否则至少有两棵子树; ④ 所有非终端结点中包含有下列信息: (n,A0,k1,A1,K2,A2,…,Kn,An) 其中,Ki (1? i ?n)为关键字,且Ki<Ki+1 (1? i <n),Ai (0? i ?n)为指向子树根结点的指针,且指针Ai(0? i<n)所 指子树中所有结点的关键字均小于Ki+1,指针An 所指子树中 的关键字均大于Kn,n(n= ?m/2? ? m-1)为关键字的个数; ⑤所有的叶子结点都出现在同一层上,并且不带信

息(可以看 作是外部结点或查失败的结点,实际上这些结点不存在,指向这 些结点的指针为空)。

3. Insertion into a B-Tree In contrast to binary search trees, B-trees are not allowed to grow at their leaves; instead, they are forced to grow at the root. General insertion method: 1. Search the tree for the new key. This search (if the key is truly new) will terminate in failure at a leaf. 2. Insert the new key into to the leaf node. If the node was not previously full, then the insertion is finished. 3. When a key is added to a full node, then the node splits into two nodes, side by side on the same level, except that the median key is not put into either of the two new nodes.

4. When a node splits, move up one level, insert the median key into this parent node, and repeat the splitting process if necessary. 5. When a key is added to a full root, then the root splits in two and the median key sent upward becomes a new root. This is the only time when the B-tree grows in height. Growth of a B-Tree The growth of a B-Tree of order 5 pg.538 Fig.11.10

4. B-Tree Declarations in C++
We add the order as a second template parameter. For example, B_tree<int, 5> sample tree; declares sample tree as a B_tree of order 5 that holds integer records. B_tree class declaration:
template <class Record, int order> class B_tree { public: // Add public methods. private: // data members B_node<Record, order> *root; // Add private auxiliary functions here. };

B-Tree Node declaration:
template <class Record, int order> struct B_node { B_node( ); private: // data members: int count; Record data[order-1]; B_node<Record, order> *branch[order]; };

记录数 记录数组 指针数组

Conventions: ?count gives the number of records in the B_node.

?If count is nonzero then the node has count+1
children. ? branch[0] points to the subtree containing all records with keys less than that in data[0]. ?For 1 ? position ? count-1, branch[position] points to the subtree with keys strictly between those in the subtrees pointed to by data[position-1] and data[position]. ?branch[count] points to the subtree with keys greater than that of data[count-1]. 参看“幻灯片25”中文定义。

5. Searching and insertion pg.539

template <class Record, int order> Error_code B_tree<Record, order>:: search_tree(Record &target) { B_node<Record, order> *sub_root=root; while(sub_root != NULL) { int position; if (search_node(sub_root, target, position) == not_present ) sub_root=sub_root->branch[position]; else { target = sub_root->data[position]; return success; } } return not_present; }

template <class Record, int order> Error_code B_tree<Record, order>::search_node ( B_node<Record, order> *current, const Record &target, int &position) {position=0; while(position<current->count && target>current->data[position]) position++; if ( position<current->count && target==current->data[position] ) return success; Perform a else return not present

; 引用参数 sequential search 带回target }

? For B-trees of large order, this function should be modified to use binary search instead of sequential search. ?Another possibility is to use a linked binary search tree instead of a sequential array of entries for each node.

through the keys 在结点中的位置

Insertion: Parameters and push down
? Insertion is done with recursion in a function called push down. ? We require that the record new_entry being inserted is not already present in the tree. The recursive function push_down uses three more output parameters. ? current is the root of the current subtree under consideration. ? If *current splits to accommodate new entry, push down returns a code of overflow, and the following come into use: ?The old node *current contains the left half of the entries.

Public Insertion Method

template <class Record, int order> Error_code B_tree<Record, order>::insert(const Record &new_entry) { Record median; B_node<Record, order> *right_branch, *new_root; Error_code result = push_down(root, new_entry, median, right_branch); if ( result==overflow) { new_root = new B_node<Record, order>; new_root->count=1; new_root->data[0]=median; new_root->branch[0]=root; new_root->branch[1]=right branch; root=new_root; result = success; The whole tree } Make grows innew_root a brand height for the whole B-tree return result; }

Recursive Insertion into a Subtree
Since we can’t insert template <class Record, int order>Error_code empty tree, the in an Search the current node. B_tree<Record, order>::push_down ( recursion terminates.

B_node<Record, order> *current, const Record &new_entry, Record &median, B_node<Record, order>* &right_branch ) { Error_code result; int position; if(current==NULL) { median=new_entry; right_branch= NULL;
result = overflow; } else { if (search_node(current, new_entry, position)==success) result=duplicate_error; else

Record extra_entry Record median and now must be added { Record extra_entry; its right branch will go up B_node<Record, order> *extra_branch; ato current to higher node

result=push_down(current->branch[position], new_entry, extra_entry, extra branch); if( result==overflow) if(current->count<order-1) { result=success; push_in(current, extra_entry, extra_branch, position); } else split_node( current, extra_entry, extra_branch, position, right_branch, median); } return result;

}

template <class Record, int order> void B_tree<Record, order>:: push_in ( B_node<Record, order> *current , const Record &entry , B_node<Record, order> *right_branch , int position )

/* Pre: current points to a node of a B_tree . The node *current is not full and entry belongs in *current at Shift all later data to the right index position . Post: entry has been inserted along with its right-hand branch right_branch into *current at index position. */ { for(int i=current->count; i>position; i--) { current->data[i] = current->data[i-1]; current->branch[i+1] = current->branch[i];

} current->data[position] = entry; current->branch[position+1] = right_branch; current->count++; }

Example of Splitting a Full Node

See pg.547 fig.11.13

Function split node, Action

template <class Record, int order> void B_tree<Record, order>::split_node ( B_node<Record, order> *current, const Record &extra_entry, B_node<Record, order> *extra_branch, int position, B_node<Record, order> * &right_half, Record &median ) /* Pre: current points to a node of a B_tree. The node *current is

Point to new_entry toright Point to index on insert subtreein node Point to new node node extra_entry to be split whereright_half of extra_entry median entry for goes (in of entries neither half)

full, but if there were room, the record extra_entry with its right-hand pointer extra_branch would belong in *current at position position ,0 ? position < order . Post: The node*current with extra_entry and pointer extra_branch at position position are divided into nodes *current and *right_half separated by a Record median.

*/

{ right_half = new B_node<Record, order>; Temporarily Second case: int mid = order/2; leave the belongs Move entries extra_entrymedian if ( position <= mid ) to right_half in left half in right_half. { for(int i=mid; i<order-1; i++) { right_half->data[i-mid] = current->data[i]; right_half->branch[i+1-mid] = current->branch[i+1]; } current->count = mid; right_half->count = order-1-mid; push_in(current, extra_entry, extra_branch, position); } else The entries from First case: { mid++; mid onentries to extra_entry belongs Move will go to for(int i=mid; i<order-1; i++) right_half in left half right_half

{ right_half->data[i-mid] = current->data[i]; right_half->branch[i+1-mid] = current->branch[i+1]; } current->count = mid; right_half->count = order-1-mid; push_in(right_half, extra_entry, extra_branch, position-mid); } median = current->data[current->count-1]; right_half->branch[0] = current->branch[current->count]; current->count--;

}

Remove median from left half

6. Deletion from a B-Tree

pg.548

?If the entry that is to be deleted is not in a leaf, then its immediate predecessor (or successor) under the natural order of keys is guaranteed to be in a leaf. ?We promote the immediate predecessor (or successor) into the position occupied by the deleted entry, and delete the entry from the leaf. ?If the leaf contains more than the minimum number of entries, then one of them can be deleted with no further action. ?If the leaf contains the minimum number, then we first look at the two leaves (or, in the case of a node on the outside,one leaf) that are immediately adjacent to each other and are children of the same node. If one of these

has more than the minimum number of entries, then one of them can be moved into the parent node, and the entry from the parent moved into the leaf where the deletion is occurring. ?If the adjacent leaf has only the minimum number of entries,then the two leaves and the median entry from the parent

can all be combined as one new leaf, which will contain no more than the maximum number of entries allowed. ?If this step leaves the parent node with too few entries, then the process propagates upward. In the limiting case, the last entry is removed from the root, and then the height of the tree decreases.

pg.549 fig.11.14

Public Deletion Method
template <class Record, int order> Error_code B tree<Record, order>:: remove(const Record &target) /*Post: If a Record with Key matching that of target belongs to the B-tree , a code of success is returned and the corresponding node is removed from the B-tree. Otherwise, a code of not_present is returned. */ { Error_code result; result = recursive_remove(root, target); if( root!=NULL && root->count==0 ) { B_node<Record, order> *old_root = root; root = root->branch[0]; delete old_root; } return result; root is now empty }

Recursive Deletion

notRemove from at a leaf node a leaf template <class Record, int order> Error_code node

B_tree<Record, order>:: recursive_remove (B_node<Record, order> *current, const Record &target) { Error_code result; int position; if( current==NULL) return not_present; if ( search_node(current,target,position)==success ) // 1 { result = success; if( current->branch[position]!=NULL ) // ? { copy_in_predecessor(current, position); recursive_remove ( current->branch[position], current->data[position] ); } // end ? else remove_data(current, position); } // end 1 The target is else result=recursive_remove( current->branch[position], in the current node target );

if( current->branch[position]!=NULL ) if( current->branch[position]->count<(order-1)/2 ) restore(current, position); return result; }

Auxiliary Functions Remove data from a leaf:
template <class Record, int order> void B_tree<Record,order>::remove_data ( B_node<Record, order> *current,int position ) { for( int i=position; i< current->count-1; i++ ) current->data[i]=current->data[i+1]; current->count--; }

Replace data by its immediate predecessor:
template <class Record, int order> void B_tree < Record, order >::copy_in_predecessor ( B_node<Record, order> *current, int position ) { B_node<Record, order> *leaf=current->branch[position]; while( leaf->branch[leaf->count]!= NULL ) leaf = leaf->branch[leaf->count]; current->data[position] = leaf->data[leaf->count-1]; }

Restore Minimum Number of Entries
(pg.552 Fig.11.15)

Function to Restore Minimum Node Entries
template <class Record, int order> void B_tree<Record, order>:: restore(B_node<Record, order> *current, int position) { if(position == current->count) if( current->branch[position-1]->count>(order-1)/2 ) move_right(current, position-1); else combine(current, position); else if( position==0 ) if(current->branch[1]->count>(order-1)/2) move_left(current, 1); else combine(current, 1); else if( current->branch[position-1]->count>(order-1)/2 ) move_right(current, position-1); else if( current->branch[position+1]->count>(order-1)/2 ) case: left case: right mov

e_left(current, position+1); remaining cases: most branch else combine(current, position);intermediate branches }

Function move_left

pg553.

Take entry Add the right-hand from the parent entry to the parent

template <class Record, int order> void B_tree<Record, order> :: move_left( B_node<Record, order> *current, int position) { B_node<Record, order> *left_branch = current->branch[position-1], *right_branch = current->branch[position]; left_branch->data[left_branch->count] = current->data[position-1]; left_branch->branch[++left_branch->count] = right_branch->branch[0]; current->data[position-1] = right_branch->data[0]; right_branch->count--;
Declare pointer

for(int i=0; i<right_branch->count; i++) { right_branch->data[i] = right_branch->data[i+1]; right_branch->branch[i] = right_branch->branch[i+1]; } right_branch->branch[right_branch->count] = right_branch->branch[right_branch->count+1]; Move right-hand entries }
to fill the hole

Function move_right

pg554.

template <class Record, int order> void B_tree<Record, order>:: move_right(B_node<Record, order> *current, int position) {B_node<Record, order> *left_branch=current->branch[position], *right_branch=current->branch[position+1];

right_branch->branch[right_branch->count+1] = right_branch->branch[right_branch->count]; for(int i=right_branch->count; i>0; i--) { right_branch->data[i] = right_branch->data[i-1]; right_branch->branch[i] = right_branch->branch[i-1]; } right_branch->count++; right_branch->data[0] = current->data[position]; right_branch->branch[0] = left_branch->branch[left_branch->count--]; current->data[position] = left_branch->data[left_branch->count]; }
Make room Take entry for newparent from entry

Function combine

pg554.

template <class Record, int order> void B_tree<Record, order>:: combine(B_node<Record, order> *current, int position) { B_node<Record, order> *left_branch = current->branch[position-1], *right_branch = current->branch[position]; left_branch->data[left_branch->count] = current->data[position-1]; left_branch->branch[++left_branch->count] = right_branch->branch[0]; for(int i=0; i<right_branch->count; i++)

{ left_branch->data[left_branch->count] = right_branch->data[i]; left_branch->branch[++left_branch->count] = right_branch->branch[i+1]; } current->count--; for(i=position-1; i<current->count; i++) { current->data[i] = current->data[i+1]; current->branch[i+1] = current->branch[i+2]; } delete right_branch; }

11.4 Red-Black Trees
1. Introduction

Red Black

Red-Black Trees as B-Trees of Order 4 See pg.557 Fig. 11.16

Red-Black Trees as B-Trees of Order 4
? Start with a B-tree of order 4, so each node contains 1, 2, or 3 entries. ? Convert a node with 3 entries into a binary search tree by:

? A node with two entries has two possible conversions:

? A node with one entry remains unchanged. ? A red-black tree is a binary search tree, with links colored red or black, obtained from a B-tree of order 4 by the above conversions. ? Searching and

traversal of a red-black tree are exactly the same as for an ordinary binary search tree. ?Insertion and deletion, require care to maintain the underlying B-tree structure.

Red-Black Trees as Binary Search Trees

? Each node of a red-black tree is colored with the same color as the link immediately above it. We thus need keep only one extra bit of information for each node to indicate its color. ? We adopt the convention that the root is colored black and all empty subtrees (corresponding to NULL links) are colored black. ?The B-tree requirement that all its empty subtrees are on the same level becomes:

The Black Condition Every simple path from the root to an empty subtree goes through the same number of black nodes.

? To guarantee that no more than three nodes are connected by red links as one B-tree node, and that nodes with three entries are a balanced binary tree, we require: The Red Condition If a node is red, then its parent exists and is black. ? We can now define: Definition A red-black tree is a binary search tree in which each node has either the color red or black and that satisfies the black and red conditions.

Analysis of Red-Black Trees Theorem 11.2 The height of a red-black tree containing n nodes is no more than 2 lg n.
?Searching a red-black tree with n nodes is O(log n) in every case. ?The time for insertion is also O(log n) [ To show this, we first need to devise the insertion algorithm. ] ? An AVL tree, in its worst case, has height about 1.44 lg n and,on average, has an even smaller height. Hence red-black trees do not achieve as good a balance as AVL trees. ?Red-black trees are not necessarily slower than AVL trees,since AVL trees may require many more rotations to maintain balance than red-black trees require.

Red-Black Tree Specification
? The red-black tree class is derived from the binary search tree class. ? We begin by incorporating colors into the nodes that will make up red-black trees:
enum Color { red, black }; template <class Record> struct RB_node : public Binary_node<Record> { Color color; RB_node(const Record &new_entry) { color=red; data=new_entry; left=right= NULL; } RB_node( ) { color = red; left = right = NULL; }

void set_color(Color c) { color = c; } Color get_color( ) const { return color; } };

? Note the inline definitions for the constructors and other methods of a red-black node.

Modified Node Specification
? To invoke get color and set color via pointers, we must add virtual functions to the base struct Binary node. ? The modified node specification is:

template <class Entry> struct Binary_node { Entry data; Binary_node<Entry> *left, *right; virtual Color get color( ) const { return red; } virtual void set color(Color c) { } Binary_node( ) { left = right = NULL; } Binary_node(const Entry &x) {data = x; left = right = NULL; } };

? With this modification, we can reuse all the methods and functions for manipulating binary search trees and their nodes.

Re

d-Black Trees Insertion ?We begin with the standard recursive algorithm for insertion into a binary search tree. The new entry will be in a new leaf node. ?The black condition requires that the new node must be red. ?If the parent of the new red node is black, then the insertion is finished, but if the parent is red, then we have introduced a violation of the red condition into the tree. ?The major work of the insertion algorithm is to remove a violation of the red condition, and we shall find several different cases that we shall need to process separately.

? We postpone this work as long as we can. When we make a node red, we do not check the conditions or repair the tree. Instead, we return from the recursive call with a status indicator showing that the node just processed is red. ? After this return, we are processing the parent node. ?If the parent is black, then the conditions for a redblack tree are satisfied and the process terminates. ?If the parent is red, then we set the status variable to show two red nodes together, linked as left child or as right child. Return from the recursive call. ?We are now processing the grandparent node. Since the root is black and the parent is red, this grandparent must exist. By the red condition, this grandparent is black since the parent was red.

? At the recursive level of the grandparent node, we transform the tree to restore the red-black conditions, using cases depending on the relative positions of the grandparent, parent, and child nodes. See following diagram.

Insertion Method Implementation

pg 561.

template <class Record> class RB_tree: public Search_tree<Record> { public: Error_code insert(const Record & new_entry); private: RB_code rb_insert ( Binary_node<Record>*&, const Record& ); RB_code modify_left ( Binary_node<Record>*&, RB_code&); RB_code modify_right ( Binary_node<Record>*&, RB_code&); };

Status indicator values:

pg 563

enum RB_code{okay, red_node, left_red, right_red, duplicate}; /* okay: The color of the current root(of the subtree) has not changed; the tree now satisfies the conditions for a redblack tree. red_node: The current root has changed from black to red; modification may or may not be needed to restore the redblack properties. right_red: The current root and its right child are now both red; a color flip or rotation is needed. left_red: The current root and its left_child are now both red; a color flip or rotation is needed. duplicate: The entry being inserted duplicates another entry; this is an error. */

Public Insertion Method

pg 563.

template <class Record> Error_code RB_tree<Record>:: insert(const Record &new_entry) /* Post: If the key ofnew_entry is already in the RB_tree , a code of duplicate error is returned. Otherwise, a code of success is returned and the Record new_entry is inserted into the tree in such a way that the properties of an RB-tree have been preserved. Uses: Methods of struct RB_node and recursive function rb_i

nsert . */

{ RB_code status = rb_insert(root, new_entry); switch (status) { case red_node: root->set_color(black); case okay: return success; case duplicate: return duplicate_error; case right_red: case left_red: cout << "WARNING: Program error in RB_tree::insert" << endl; return internal_error; } Always split Convert private }

See pg. 562 Fig. 11.17

the root node RB_code to to Error_code. publickeep it black

Recursive Insertion Function

pg 563.

template <class Record> RB_code RB_tree<Record>:: rb_insert ( Binary_node<Record>* &current, const Record &new_entry) /* Pre: current is either NULL or points to the first node of a subtree of an RB_tree Post: If the key of new_entry is already in the subtree, a code of duplicate is returned. Otherwise, the Record new_entry is inserted into the subtree pointed to by current. The properties of a red-black tree have been restored, except possibly at the root current and one of its children, whose status is given by the output RB_code. */

{ RB_code status, child_status; if( current==NULL ) { current = new RB_node<Record>(new_entry); status = red_node; } else if( new_entry==current->data ) return duplicate; else if( new_entry<current->data) { child_status = rb_insert(current->left, new_entry); status = modify_left(current, child_status); } else { child_status = rb_insert(current->right, new_entry); status = modify_right(current, child_status); } return status; }

Checking the Status: Left Child

pg 563.

template <class Record> RB_code RB_tree<Record>:: modify_left ( Binary_node<Record>* &current, RB_code &child_status) /* Pre: An insertion has been made in the left_subtree of *current that has returned the value of child_status for this subtree. Post: Any color change or rotation needed for the tree rooted at current has been made, and a status code is returned. Uses: Methods of struct RB_node, with rotate_right , double_rotate_right, and flip_color. */

{ RB_code status = okay; Binary_node<Record> *aunt = current->right; Color aunt_color = black; if(aunt != NULL) aunt_color = aunt->get_color( ); switch (child_status) { case okay: break; case red_node: if(current->get_color()==red) status = left_red; else status = okay; break; case left_red: if( aunt_color==black ) status = rotate_right(current); else status = flip_color(current); action needed, No current is black, break; as tree is already OK
left is red, so OK

case right_red: if( aunt_color==black ) status = double rotate_right(current); else status = flip_color(current); break; } return status; }

Red-Black Trees

Insert 15 in Red-Black Trees


上一篇:弟子规 带拼音2-1
下一篇:蓝莓鳕鱼
网站首页网站地图 站长统计
All rights reserved Powered by 海文库
copyright ©right 2010-2011。
文档资料库内容来自网络,如有侵犯请联系客服。zhit326@126.com