java 轉換中文數字為阿拉伯數字
可以轉換以下格式的數字:
一百
三千五
二萬三千五百三
3千4百
package com.baijob.vsearch.util;
import java.util.*;
/**
* 處理數詞的工具類
* @author liushiquan
*
*/
public class DigitUtil {
/**
* 阿拉伯數字
*/
private static Set<Character> araDigits = new HashSet<Character>();
/**
* 漢字中的數字字符
*/
private static char[] SCDigits = {'零','一','二','三','四','五','六','七','八','九','十','百','千','萬','億'};
/**
* 漢字中的大寫數字字符
*/
private static char[] TCDigits = {'零','壹','貳','叁','肆','伍','陸','柒','捌','玖','拾','佰','仟','萬','億'};
/**
* 繁體中文和簡體中文的對應關系
*/
private static Map<Character,Character> map = new HashMap<Character,Character>();
static {
for (int i = 0; i < TCDigits.length; i++) {
map.put(TCDigits[i], SCDigits[i]);
}
for (char i = '0'; i <= '9'; i++) {
araDigits.add(i);
}
}
private DigitUtil(){
}
public static void main(String[] args) {
System.out.println(parseDigits("零三"));
}
/**
* 解析中文格式的數字,假定參數中全是漢字,否則會解析異常,解析失敗返回null
* @param hanzi
* @return
*/
public static Integer parseDigits(String hanzi) {
if (!isDigits(hanzi))
return null;
int ret;
try {
if (hanzi.charAt(0) == '+')
hanzi = hanzi.substring(1);
ret = Integer.parseInt(hanzi);
} catch (Exception e) {
char[] chars = hanzi.toCharArray();
changeTCtoSC(chars);
ret = parse(chars,0,chars.length,1);
}
return ret;
}
public static boolean isDigits(String s) {
if (s.charAt(0) == '+')
s = s.substring(1);
try {
Integer.parseInt(s);
return true;
} catch (Exception e) {
for (int i = 0; i < s.length(); i++) {
char c = s.charAt(i);
if (!map.values().contains(c) && !araDigits.contains(c))
return false;
}
return true;
}
}
private static int parse(char[] chars,int start,int end, int preNumber) {
int ret = 0;
if (start == end) {
ret = 0;
} else if (start + 1 == end) {
switch (chars[start]) {
case '一':
case '1':
ret = 1 * preNumber;
break;
case '二':
case '2':
ret = 2 * preNumber;
break;
case '三':
case '3':
ret = 3 * preNumber;
break;
case '四':
case '4':
ret = 4 * preNumber;
break;
case '五':
case '5':
ret = 5 * preNumber;
break;
case '六':
case '6':
ret = 6 * preNumber;
break;
case '七':
case '7':
ret = 7 * preNumber;
break;
case '八':
case '8':
ret = 8 * preNumber;
break;
case '九':
case '9':
ret = 9 * preNumber;
break;
}
} else {
int index;
if ((index = indexOf(chars,start,end,'零')) == 0 || (index = indexOf(chars,start,end,'0')) == 0) {
ret = parse(chars, start + 1, end, 1);
} else if ((index = indexOf(chars,start,end,'億')) != -1) {
ret = parse(chars, start,index, 1) * 100000000 + parse(chars,index + 1,end,10000000);
} else if ((index = indexOf(chars,start,end,'萬')) != -1) {
ret = parse(chars, start,index, 1) * 10000 + parse(chars,index + 1,end,1000);
} else if ((index = indexOf(chars,start,end,'千')) != -1) {
ret = parse(chars, start, index, 1) * 1000 + parse(chars,index + 1,end,100);
} else if ((index = indexOf(chars,start,end,'百')) != -1) {
ret = parse(chars, start, index, 1) * 100 + parse(chars,index + 1,end,10);
} else if ((index = indexOf(chars,start,end,'十')) != -1) {
ret = parse(chars, start, index, 1) * 10 + parse(chars,index + 1,end,1);
}
}
return ret;
}
private static int indexOf(char[] chars, int start, int end, char c) {
for (int i = start; i < end; i++) {
if (chars[i] == c)
return i;
}
return -1;
}
/**
* 將繁體中文轉換為簡體中文
* @param chars
*/
private static void changeTCtoSC(char[] chars) {
for (int i = 0; i < chars.length; i++) {
Character c = map.get(chars[i]);
if (c != null)
chars[i] = c;
}
}
}
//該片段來自于http://outofmemory.cn
本文由用戶 ErvCavill 自行上傳分享,僅供網友學習交流。所有權歸原作者,若您的權利被侵害,請聯系管理員。
轉載本站原創文章,請注明出處,并保留原始鏈接、圖片水印。
本站是一個以用戶分享為主的開源技術平臺,歡迎各類分享!