`

java汉字转拼音或首字母,手写版非pinyin4j,可排除例外

    博客分类:
  • JAVA
阅读更多
package com.anxin.ssk.common;

import java.io.UnsupportedEncodingException;
import java.util.HashMap;
import java.util.Map;
import java.util.Map.Entry;

/**
 * 取得给定汉字串的首字母串,即声母串 Title: ChineseCharToEn
 * 
 * @author: (le.qiao)
 * @e-mail: qiaolevip@gmail.com
 * @myblog: <a href="http://qiaolevip.iteye.com">http://qiaolevip.iteye.com</a>
 * @date: 2014-1-15 注:只支持GB2312字符集中的汉字
 * 
 */
public class LetterUtil {

	private final static int[] li_SecPosValue = { 1601, 1637, 1833, 2078, 2274, 2302, 2433, 2594, 2787, 3106, 3212, 3472, 3635, 3722, 3730, 3858, 4027, 4086,
			4390, 4558, 4684, 4925, 5249, 5590 };

	private final static String[] lc_FirstLetter = { "a", "b", "c", "d", "e", "f", "g", "h", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "w", "x",
			"y", "z" };

	/**
	 * 取得给定汉字串的首字母串,即声母串
	 * 
	 * @param str 给定汉字串
	 * @return 声母串
	 */
	public static String getAllFirstLetter(String str) {
		if (str == null || str.trim().length() == 0) {
			return "";
		}

		String _str = "";
		for (int i = 0; i < str.length(); i++) {
			_str = _str + getFirstLetter(str.substring(i, i + 1));
		}

		return _str;
	}

	/**
	 * 取得给定汉字的首字母,即声母
	 * 
	 * @param chinese 给定的汉字
	 * @return 给定汉字的声母
	 */
	public static String getFirstLetter(String chinese) {
		if (chinese == null || chinese.trim().length() == 0) {
			return "";
		}
		chinese = conversionStr(chinese, "GB2312", "ISO8859-1");

		if (chinese.length() > 1) // 判断是不是汉字
		{
			int li_SectorCode = (int) chinese.charAt(0); // 汉字区码
			int li_PositionCode = (int) chinese.charAt(1); // 汉字位码
			li_SectorCode = li_SectorCode - 160;
			li_PositionCode = li_PositionCode - 160;
			int li_SecPosCode = li_SectorCode * 100 + li_PositionCode; // 汉字区位码
			if (li_SecPosCode > 1600 && li_SecPosCode < 5590) {
				for (int i = 0; i < 23; i++) {
					if (li_SecPosCode >= li_SecPosValue[i] && li_SecPosCode < li_SecPosValue[i + 1]) {
						chinese = lc_FirstLetter[i];
						break;
					}
				}
			} else // 非汉字字符,如图形符号或ASCII码
			{
				chinese = conversionStr(chinese, "ISO8859-1", "GB2312");
				chinese = chinese.substring(0, 1);
				Map<String, String> exceptWords = new HashMap<String, String>();
// 欢迎大家补充说明~欢迎大家补充说明~欢迎大家补充说明
				exceptWords.put("a", "庵");
				exceptWords.put("g", "崮");
				exceptWords.put("l", "崂");
				exceptWords.put("j", "泾");
				exceptWords.put("r", "榕");
				exceptWords.put("x", "鑫");
				exceptWords.put("y", "懿");
				exceptWords.put("z", "梓");

				for (Entry<String, String> letterSet : exceptWords.entrySet()) {
					if (letterSet.getValue().indexOf(chinese) != -1) {
						chinese = letterSet.getKey();
						break;
					}
				}
			}
		}

		return chinese;
	}

	/**
	 * 字符串编码转换
	 * 
	 * @param str 要转换编码的字符串
	 * @param charsetName 原来的编码
	 * @param toCharsetName 转换后的编码
	 * @return 经过编码转换后的字符串
	 */
	private static String conversionStr(String str, String charsetName, String toCharsetName) {
		try {
			str = new String(str.getBytes(charsetName), toCharsetName);
		} catch (UnsupportedEncodingException ex) {
			System.out.println("字符串编码转换异常:" + ex.getMessage());
		}
		return str;
	}

	public static void main(String[] args) {
		System.out.println("获取拼音首字母:" + LetterUtil.getAllFirstLetter("泾崂"));
		String address = "(金浜小区)栖山路1689弄";
		address = address.substring(address.indexOf(")") + 1);
		System.out.println("获取拼音首字母:" + LetterUtil.getFirstLetter(address));
	}
}

 

分享到:
评论

相关推荐

Global site tag (gtag.js) - Google Analytics