简介
Trie树,又称为前缀树或字典树,是一种有序树,用于保存关联数组,其中的键通常是字符串。与二叉查找树不同,键不是直接保存在节点中,而是由节点在树中的位置决定。一个节点的所有子孙都有相同的前缀,也就是这个节点对应的字符串,而根节点对应空字符串。
它的主要特点如下:
根节点不包含字符,除根节点外的每一个节点都只包含一个字符。
从根节点到某一节点,路径上经过的字符连接起来,为该节点对应的字符串。
每个节点的所有子节点包含的字符都不相同。
如下是一棵典型的Trie树:
Trie的来源是Retrieval,它常用于前缀匹配和词频统计。可能有人要说了,词频统计简单啊,一个hash或者一个堆就可以搞定,但问题来了,如果内存有限呢?还能这么 玩吗?所以这里我们就可以用trie树来压缩下空间,因为公共前缀都是用一个节点保存的。
1、定义
这里为了简化,只考虑了26个小写字母。
首先是节点的定义:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
|
public class TrieNode { public TrieNode[] children; public char data; public int freq; public TrieNode() { //因为有26个字母 children = new TrieNode[ 26 ]; freq = 0 ; } } |
然后是Trie树的定义:
1
2
3
4
5
6
7
8
9
10
11
|
public class TrieTree { private TrieNode root; public TrieTree(){ root= new TrieNode(); } ... } |
2、插入
由于是26叉树,故可通过charArray[index]-‘a';来得知字符应该放在哪个孩子中。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
|
public void insert(String word){ if (TextUtils.isEmpty(word)){ return ; } insertNode(root,word.toCharArray(), 0 ); } private static void insertNode(TrieNode rootNode, char []charArray, int index){ int k=charArray[index]- 'a' ; if (k< 0 ||k> 25 ){ throw new RuntimeException( "charArray[index] is not a alphabet!" ); } if (rootNode.children[k]== null ){ rootNode.children[k]= new TrieNode(); rootNode.children[k].data=charArray[index]; } if (index==charArray.length- 1 ){ rootNode.children[k].freq++; return ; } else { insertNode(rootNode.children[k],charArray,index+ 1 ); } } |
3、移除节点
移除操作中,需要对词频进行减一操作。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
|
public void remove(String word){ if (TextUtils.isEmpty(word)){ return ; } remove(root,word.toCharArray(), 0 ); } private static void remove(TrieNode rootNode, char []charArray, int index){ int k=charArray[index]- 'a' ; if (k< 0 ||k> 25 ){ throw new RuntimeException( "charArray[index] is not a alphabet!" ); } if (rootNode.children[k]== null ){ //it means we cannot find the word in this tree return ; } if (index==charArray.length- 1 &&rootNode.children[k].freq > 0 ){ rootNode.children[k].freq--; } remove(rootNode.children[k],charArray,index+ 1 ); } |
4、查找频率
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
|
public int getFreq(String word){ if (TextUtils.isEmpty(word)){ return 0 ; } return getFreq(root,word.toCharArray(), 0 ); } private static int getFreq(TrieNode rootNode, char []charArray, int index){ int k=charArray[index]- 'a' ; if (k< 0 ||k> 25 ){ throw new RuntimeException( "charArray[index] is not a alphabet!" ); } //it means the word is not in the tree if (rootNode.children[k]== null ){ return 0 ; } if (index==charArray.length- 1 ){ return rootNode.children[k].freq; } return getFreq(rootNode.children[k],charArray,index+ 1 ); } |
5、测试
测试代码如下:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
public static void test(){ TrieTree trieTree= new TrieTree(); String sourceStr= "Democratic presumptive nominee Hillary Clintons campaign posed pounced on Trumps assertion that British term monetary turmoil might benefit his business venture in Scotland" ; //String sourceStr="the that"; sourceStr=sourceStr.toLowerCase(); String[]strArray=sourceStr.split( " " ); for (String str:strArray){ trieTree.insert(str); } String sourceStr2= "Every president is tested by world events But Donald Trump thinks about how is his golf resort can profit from that" ; sourceStr2=sourceStr2.toLowerCase(); String[]strArray2=sourceStr2.split( " " ); for (String str:strArray2){ trieTree.insert(str); } BinaryTree.print( "frequence of 'that':" +trieTree.getFreq( "that" )); BinaryTree.print( "\nfrequence of 'donald':" +trieTree.getFreq( "donald" )); trieTree.remove( "that" ); BinaryTree.print( "\nafter remove 'that' once,freq of 'that':" +trieTree.getFreq( "that" )); trieTree.remove( "that" ); BinaryTree.print( "\nafter remove 'that' twice,freq of 'that':" +trieTree.getFreq( "that" )); trieTree.remove( "donald" ); BinaryTree.print( "\nafter remove 'donald' once,freq of 'donald':" +trieTree.getFreq( "donald" )); BinaryTree.reallyStartPrint(); } |
测试结果如下:
总结
以上就是这篇文章的全部内容了,希望本文的内容对大家的学习或者工作能带来一定的帮助,如果有疑问大家可以留言交流。