转换中文数字为阿拉伯数字

可以转换以下格式的数字：
一百
三千五
二万三千五百三
3千4百
  1 package com.baijob.vsearch.util;
  2 
  3 import java.util.*;
  4 
  5 /**
  6  * 处理数词的工具类
  7  * @author liushiquan
  8  *
  9  */
 10 public class DigitUtil {
 11     /**
 12      * 阿拉伯数字
 13      */
 14     private static Set<Character> araDigits = new HashSet<Character>();
 15     /**
 16      * 汉字中的数字字符
 17      */
 18     private static char[] SCDigits = {'零','一','二','三','四','五','六','七','八','九','十','百','千','万','亿'};
 19 
 20     /**
 21      * 汉字中的大写数字字符
 22      */
 23     private static char[] TCDigits = {'零','壹','贰','叁','肆','伍','陆','柒','捌','玖','拾','佰','仟','万','亿'};
 24     /**
 25      * 繁体中文和简体中文的对应关系
 26      */
 27     private static Map<Character,Character> map = new HashMap<Character,Character>();
 28     static {
 29         for (int i = 0; i < TCDigits.length; i++) {
 30             map.put(TCDigits[i], SCDigits[i]);
 31         }
 32         for (char i = '0'; i <= '9'; i++) {
 33             araDigits.add(i);
 34         }
 35     }
 36     private DigitUtil(){
 37 
 38     }
 39     public static void main(String[] args) {
 40         System.out.println(parseDigits("零三"));
 41     }
 42     /**
 43      * 解析中文格式的数字，假定参数中全是汉字，否则会解析异常，解析失败返回null
 44      * @param hanzi
 45      * @return
 46      */
 47     public static Integer parseDigits(String hanzi) {
 48         if (!isDigits(hanzi))
 49             return null;
 50         int ret;
 51         try {
 52             if (hanzi.charAt(0) == '+')
 53                 hanzi = hanzi.substring(1);
 54 
 55             ret = Integer.parseInt(hanzi);
 56         } catch (Exception e) {
 57 
 58             char[] chars = hanzi.toCharArray();
 59             changeTCtoSC(chars);
 60 
 61             ret = parse(chars,0,chars.length,1);
 62         }
 63 
 64         return ret;
 65     }
 66     public static boolean isDigits(String s) {
 67         if (s.charAt(0) == '+')
 68             s = s.substring(1);
 69         try {
 70             Integer.parseInt(s);
 71             return true;
 72         } catch (Exception e) {
 73             for (int i = 0; i < s.length(); i++) {
 74                 char c = s.charAt(i);
 75                 if (!map.values().contains(c) && !araDigits.contains(c))
 76                     return false;
 77             }
 78 
 79             return true;
 80         }
 81     }
 82     private static int parse(char[] chars,int start,int end, int preNumber) {
 83         int ret = 0;
 84         if (start == end) {
 85             ret = 0;
 86         } else if (start + 1 == end) {
 87             switch (chars[start]) {
 88             case '一':
 89             case '1':
 90                 ret = 1 * preNumber;
 91                 break;
 92             case '二':
 93             case '2':
 94                 ret = 2 * preNumber;
 95                 break;
 96             case '三':
 97             case '3':
 98                 ret = 3 * preNumber;
 99                 break;
100             case '四':
101             case '4':
102                 ret = 4 * preNumber;
103                 break;
104             case '五':
105             case '5':
106                 ret = 5 * preNumber;
107                 break;
108             case '六':
109             case '6':
110                 ret = 6 * preNumber;
111                 break;
112             case '七':
113             case '7':
114                 ret = 7 * preNumber;
115                 break;
116             case '八':
117             case '8':
118                 ret = 8 * preNumber;
119                 break;
120             case '九':
121             case '9':
122                 ret = 9 * preNumber;
123                 break;
124             }
125         } else {
126             int index;
127             if ((index = indexOf(chars,start,end,'零')) == 0 || (index = indexOf(chars,start,end,'0')) == 0) {
128                 ret = parse(chars, start + 1, end, 1);
129             } else if ((index = indexOf(chars,start,end,'亿')) != -1) {
130                 ret = parse(chars, start,index, 1) * 100000000 + parse(chars,index + 1,end,10000000);
131             } else if ((index = indexOf(chars,start,end,'万')) != -1) {
132                 ret = parse(chars, start,index, 1) * 10000 + parse(chars,index + 1,end,1000);
133             } else if ((index = indexOf(chars,start,end,'千')) != -1) {
134                 ret = parse(chars, start, index, 1) * 1000 + parse(chars,index + 1,end,100);
135             } else if ((index = indexOf(chars,start,end,'百')) != -1) {
136                 ret = parse(chars, start, index, 1) * 100 + parse(chars,index + 1,end,10);
137             } else if ((index = indexOf(chars,start,end,'十')) != -1) {
138                 ret = parse(chars, start, index, 1) * 10 + parse(chars,index + 1,end,1);
139             }
140 
141         }
142         return ret;
143     }
144     private static int indexOf(char[] chars, int start, int end, char c) {
145         for (int i = start; i < end; i++) {
146             if (chars[i] == c)
147                 return i;
148         }
149         return -1;
150     }
151     /**
152      * 将繁体中文转换为简体中文
153      * @param chars
154      */
155     private static void changeTCtoSC(char[] chars) {
156         for (int i = 0; i < chars.length; i++) {
157             Character c = map.get(chars[i]);
158             if (c != null)
159                 chars[i] = c;
160         }
161     }
162 }
163 //该片段来自于http://outofmemory.cn
posted @ 2015-06-16 14:43 在北京的四川小伙阅读(538) 评论(0) 编辑收藏举报
刷新页面返回顶部