C#将汉字转换为拼音首字母
关于这个话题以前曾经使用过一个简便的算法很长时间, 代码如下:
private string ToPinyinSingle(string str)
{
if (str.CompareTo("吖") < 0)
return str;
if (str.CompareTo("八") < 0)
return "a";
if (str.CompareTo("嚓") < 0)
return "b";
if (str.CompareTo("咑") < 0)
return "c";
if (str.CompareTo("妸") < 0)
return "d";
if (str.CompareTo("发") < 0)
return "e";
if (str.CompareTo("旮") < 0)
return "f";
if (str.CompareTo("铪") < 0)
return "g";
if (str.CompareTo("讥") < 0)
return "h";
if (str.CompareTo("咔") < 0)
return "j";
if (str.CompareTo("垃") < 0)
return "k";
if (str.CompareTo("嘸") < 0)
return "l";
if (str.CompareTo("拏") < 0)
return "m";
if (str.CompareTo("噢") < 0)
return "n";
if (str.CompareTo("妑") < 0)
return "o";
if (str.CompareTo("七") < 0)
return "p";
if (str.CompareTo("亽") < 0)
return "q";
if (str.CompareTo("仨") < 0)
return "r";
if (str.CompareTo("他") < 0)
return "s";
if (str.CompareTo("哇") < 0)
return "t";
if (str.CompareTo("夕") < 0)
return "w";
if (str.CompareTo("丫") < 0)
return "x";
if (str.CompareTo("帀") < 0)
return "y";
if (str.CompareTo("咗") < 0)
return "z";
return str;
}
这个函数只处理单个汉字, 简单地加个循环就可以让它处理文字串了.
在.net 3.5下, 它一直工作得很好, 虽然偶尔也有出错的时候, 但是概率极低, 基本上可以忽略不计.
然而后来我把项目升级到.net 4.0以后, 发现出错的几率直线上升, 已经高得无法容忍的程度了(例如, "梅" 会返回"L"), 简单查了一下, 没找到微软关于String.CompareTo函数有什么变化的说明, 束手无策, 于是换用另一个也很简单的算法(http://topic.csdn.net/u/20090219/12/61745e3a-a39e-4f4d-8985-67d124236694.html):
static public string getSpell(string cn)
{
byte[] arrCN = System.Text.Encoding.Default.GetBytes(cn);
if(arrCN.Length > 1)
{
int area = (short)arrCN[0];
int pos = (short)arrCN[1];
int code = (area<<8) + pos;
int[] areacode = {45217,45253,45761,46318,46826,47010,47297,47614,48119,48119,49062,49324,49896,50371,50614,50622,50906,51387,51446,52218,52698,52698,52698,52980,53689,54481};
for(int i=0;i<26;i++)
{
int max = 55290;
if(i != 25) max = areacode[i+1];
if(areacode[i]<=code && code<max)
{
return System.Text.Encoding.Default.GetString(new byte[]{(byte)(65+i)});
}
}
return "?";
}
else return cn;
}
但是这个函数出错的概率也很高, 例如"闫""窦""圳" 等都无法识别, 追查了一下原因, 发现原来对GB2312编码来说, 存放规定是这样的:
01-09区为特殊符号。
16-55区为一级汉字，按拼音排序。
56-87区为二级汉字，按部首/笔画排序。
每个汉字及符号以两个字节来表示。第一个字节称为“高位字节”，第二个字节称为“低位字节”。
“高位字节”使用了0xA1-0xF7(把01-87区的区号加上0xA0)，“低位字节”使用了0xA1-0xFE(把01-94加上0xA0)。
例如“啊”字在大多数程序中，会以0xB0A1储存。（与区位码对比：0xB0=0xA0+16,0xA1=0xA0+1）
上述几个字位置码都大于55290, 显然是二级汉字, 这个算法就处理不了了, 换言之, 这种写法只能用于处理一级汉字. 这当然是不可接受的.
后来翻查良久, 终于找到一个用C++写的算法, 可以同时处理一级汉字和二级汉字(http://download.csdn.net/detail/ronjay/1955072), 我把它改写成了C#, 代码如下:
public class ChineseToPinYin
{
#region " 全局变量 "
private static string[] _regionChar = new string[32]
{
"CJWGNSPGCGNESYPBTYYZDXYKYGTDJNNJQMBSGZSCYJSYYQPGKBZGYCYWJKGKLJSWKPJQHYTWDDZLSGMRYPYWWCCKZNKYDG",
"TTNGJEYKKZYTCJNMCYLQLYPYQFQRPZSLWBTGKJFYXJWZLTBNCXJJJJZXDTTSQZYCDXXHGCKBPHFFSSWYBGMXLPBYLLLHLX",
"SPZMYJHSOJNGHDZQYKLGJHSGQZHXQGKEZZWYSCSCJXYEYXADZPMDSSMZJZQJYZCDJZWQJBDZBXGZNZCPWHKXHQKMWFBPBY",
"DTJZZKQHYLYGXFPTYJYYZPSZLFCHMQSHGMXXSXJJSDCSBBQBEFSJYHWWGZKPYLQBGLDLCCTNMAYDDKSSNGYCSGXLYZAYBN",
"PTSDKDYLHGYMYLCXPYCJNDQJWXQXFYYFJLEJBZRXCCQWQQSBNKYMGPLBMJRQCFLNYMYQMSQTRBCJTHZTQFRXQHXMJJCJLX",
"QGJMSHZKBSWYEMYLTXFSYDSGLYCJQXSJNQBSCTYHBFTDCYZDJWYGHQFRXWCKQKXEBPTLPXJZSRMEBWHJLBJSLYYSMDXLCL",
"QKXLHXJRZJMFQHXHWYWSBHTRXXGLHQHFNMNYKLDYXZPWLGGTMTCFPAJJZYLJTYANJGBJPLQGDZYQYAXBKYSECJSZNSLYZH",
"ZXLZCGHPXZHZNYTDSBCJKDLZAYFMYDLEBBGQYZKXGLDNDNYSKJSHDLYXBCGHXYPKDJMMZNGMMCLGWZSZXZJFZNMLZZTHCS",
"YDBDLLSCDDNLKJYKJSYCJLKOHQASDKNHCSGANHDAASHTCPLCPQYBSDMPJLPCJOQLCDHJJYSPRCHNWJNLHLYYQYYWZPTCZG",
"WWMZFFJQQQQYXACLBHKDJXDGMMYDJXZLLSYGXGKJRYWZWYCLZMSSJZLDBYDCFCXYHLXCHYZJQSFQAGMNYXPFRKSSBJLYXY",
"SYGLNSCMHCWWMNZJJLXXHCHSYDSTTXRYCYXBYHCSMXJSZNPWGPXXTAYBGAJCXLYSDCCWZOCWKCCSBNHCPDYZNFCYYTYCKX",
"KYBSQKKYTQQXFCWCHCYKELZQBSQYJQCCLMTHSYWHMKTLKJLYCXWHEQQHTQHZPQSQSCFYMMDMGBWHWLGSSLYSDLMLXPTHMJ",
"HWLJZYHZJXHTXJLHXRSWLWZJCBXMHZQXSDZPMGFCSGLSXYMJSHXPJXWMYQKSMYPLRTHBXFTPMHYXLCHLHLZYLXGSSSSTCL",
"SLDCLRPBHZHXYYFHBBGDMYCNQQWLQHJJZYWJZYEJJDHPBLQXTQKWHLCHQXAGTLXLJXMSLXHTZKZJECXJCJNMFBYCSFYWYB",
"JZGNYSDZSQYRSLJPCLPWXSDWEJBJCBCNAYTWGMPAPCLYQPCLZXSBNMSGGFNZJJBZSFZYNDXHPLQKZCZWALSBCCJXJYZGWK",
"YPSGXFZFCDKHJGXDLQFSGDSLQWZKXTMHSBGZMJZRGLYJBPMLMSXLZJQQHZYJCZYDJWBMJKLDDPMJEGXYHYLXHLQYQHKYCW",
"CJMYYXNATJHYCCXZPCQLBZWWYTWBQCMLPMYRJCCCXFPZNZZLJPLXXYZTZLGDLDCKLYRZZGQTGJHHHJLJAXFGFJZSLCFDQZ",
"LCLGJDJCSNCLLJPJQDCCLCJXMYZFTSXGCGSBRZXJQQCTZHGYQTJQQLZXJYLYLBCYAMCSTYLPDJBYREGKLZYZHLYSZQLZNW",
"CZCLLWJQJJJKDGJZOLBBZPPGLGHTGZXYGHZMYCNQSYCYHBHGXKAMTXYXNBSKYZZGJZLQJDFCJXDYGJQJJPMGWGJJJPKQSB",
"GBMMCJSSCLPQPDXCDYYKYFCJDDYYGYWRHJRTGZNYQLDKLJSZZGZQZJGDYKSHPZMTLCPWNJAFYZDJCNMWESCYGLBTZCGMSS",
"LLYXQSXSBSJSBBSGGHFJLWPMZJNLYYWDQSHZXTYYWHMCYHYWDBXBTLMSYYYFSXJCSDXXLHJHFSSXZQHFZMZCZTQCXZXRTT",
"DJHNNYZQQMNQDMMGYYDXMJGDHCDYZBFFALLZTDLTFXMXQZDNGWQDBDCZJDXBZGSQQDDJCMBKZFFXMKDMDSYYSZCMLJDSYN",
"SPRSKMKMPCKLGDBQTFZSWTFGGLYPLLJZHGJJGYPZLTCSMCNBTJBQFKTHBYZGKPBBYMTTSSXTBNPDKLEYCJNYCDYKZDDHQH",
"SDZSCTARLLTKZLGECLLKJLQJAQNBDKKGHPJTZQKSECSHALQFMMGJNLYJBBTMLYZXDCJPLDLPCQDHZYCBZSCZBZMSLJFLKR",
"ZJSNFRGJHXPDHYJYBZGDLQCSEZGXLBLGYXTWMABCHECMWYJYZLLJJYHLGBDJLSLYGKDZPZXJYYZLWCXSZFGWYYDLYHCLJS",
"CMBJHBLYZLYCBLYDPDQYSXQZBYTDKYXJYYCNRJMPDJGKLCLJBCTBJDDBBLBLCZQRPPXJCGLZCSHLTOLJNMDDDLNGKAQHQH",
"JGYKHEZNMSHRPHQQJCHGMFPRXHJGDYCHGHLYRZQLCYQJNZSQTKQJYMSZSWLCFQQQXYFGGYPTQWLMCRNFKKFSYYLQBMQAMM",
"MYXCTPSHCPTXXZZSMPHPSHMCLMLDQFYQXSZYJDJJZZHQPDSZGLSTJBCKBXYQZJSGPSXQZQZRQTBDKYXZKHHGFLBCSMDLDG",
"DZDBLZYYCXNNCSYBZBFGLZZXSWMSCCMQNJQSBDQSJTXXMBLTXZCLZSHZCXRQJGJYLXZFJPHYMZQQYDFQJJLZZNZJCDGZYG",
"CTXMZYSCTLKPHTXHTLBJXJLXSCDQXCBBTJFQZFSLTJBTKQBXXJJLJCHCZDBZJDCZJDCPRNPQCJPFCZLCLZXZDMXMPHJSGZ",
"GSZZQJYLWTJPFSYASMCJBTZKYCWMYTCSJJLJCQLWZMALBXYFBPNLSFHTGJWEJJXXGLLJSTGSHJQLZFKCGNNDSZFDEQFHBS",
"AQTGLLBXMMYGSZLDYDQMJJRGBJTKGDHGKBLQKBDMBYLXWCXYTTYBKMRTJZXQJBHLMHMJJZMQASLDCYXYQDLQCAFYWYXQHZ"
};
private static System.Text.Encoding _encoding = System.Text.Encoding.GetEncoding("GB2312");
#endregion
private static bool In(int lp, int hp, int value)
{
return ((value <= hp) && (value >= lp));
}
public static char GetFirstChar(string chineseChar)
{
var bytes = _encoding.GetBytes(chineseChar);
if (bytes.Length != 2)
return chineseChar[0];
return GetChar(bytes[0], bytes[1], chineseChar);
}
private static char GetChar(byte c1, byte c2, string originChar)
{
var Hi = c1 << 8;
var Lo = c2;
int n = Hi + Lo;
if (n <= 0xD7F9)
{
if (In(0xB0A1, 0xB0C4, n)) return 'A';
if (In(0XB0C5, 0XB2C0, n)) return 'B';
if (In(0xB2C1, 0xB4ED, n)) return 'C';
if (In(0xB4EE, 0xB6E9, n)) return 'D';
if (In(0xB6EA, 0xB7A1, n)) return 'E';
if (In(0xB7A2, 0xB8C0, n)) return 'F';
if (In(0xB8C1, 0xB9FD, n)) return 'G';
if (In(0xB9FE, 0xBBF6, n)) return 'H';
if (In(0xBBF7, 0xBFA5, n)) return 'J';
if (In(0xBFA6, 0xC0AB, n)) return 'K';
if (In(0xC0AC, 0xC2E7, n)) return 'L';
if (In(0xC2E8, 0xC4C2, n)) return 'M';
if (In(0xC4C3, 0xC5B5, n)) return 'N';
if (In(0xC5B6, 0xC5BD, n)) return 'O';
if (In(0xC5BE, 0xC6D9, n)) return 'P';
if (In(0xC6D1, 0xC8BA, n)) return 'Q';
if (In(0xC8BB, 0xC8F5, n)) return 'R';
if (In(0xC8F6, 0xCBF9, n)) return 'S';
if (In(0xCBFA, 0xCDD9, n)) return 'T';
if (In(0xCDDA, 0xCEF3, n)) return 'W';
if (In(0xCEF4, 0xD1B8, n)) return 'X';
if (In(0xD1B9, 0xD4D0, n)) return 'Y';
if (In(0xD4D1, 0xD7F9, n)) return 'Z';
return originChar[0];
}
else
{
var b1 = (c1 & 0x7F) - 0x20 - 56;
var b2 = (c2 & 0x7F) - 0x20 - 1;
if (b1 >= 0 && b1 <= 31 && b2 >= 0 && b2 <= 93)
{
return _regionChar[b1][b2];
}
return originChar[0];
}
}
}
这个算法目前还没有发现哪个汉字会出错.