事实证明，网上的很多东西都是扯淡，不过原理还是有的，我碰到一个需求，给定一段文字中（GBK编码），去除所有符号，不相干符号，包括标点，日文片假名之类的，于是我在网上搜啊，搜啊，得到的结果要么就是错的，要么就根本好不相干，被抄袭得最厉害的是http://community.mybbchina.net/thread-212-post-507.html，可怜的作者，我也不知道这些人为什么抄这个，我就没测试成功过，连GBK的字符集区间都没弄清楚。
其实指定字符编码匹配中文很简单，GBK编码表中很明确的规定了所有的符号位置，而且刚好是一个区间，在维基百科上有详细的介绍，比如我要把类似
addd中华ds,#￥%…&（（212））｝｝Ａ■g民国ds中-啊国fjsd,【】啊Ｙ
这样的字符串中所有非中文、英文、数字字符全部去除掉，姑且叫去除的字符为火星文，去除火星文的PHP写法为:
preg_replace("/([\xA1-\xA9].{1}?)*?/", "", $str);
这样所有在全角状态下输入的非中文，英文和数字将被替换掉，如果还需要进一步处理英文状态下的，那就判断一下ASCII码值吧，或者你怕麻烦也可以把所有的都列出来，然后替换,可以参考下面，在网上找来的，非我写的，忘了出处了
/** * 清除所有常见标点符号 * @param $pointer * @return unknown_type */ function clear_point($pointer) { return str_replace ( array(' ',"~","!","@","#","$","%","^","&","*",",",".","?",";",":","'",'"'," [", "]","{","}","！"," ￥","……","…","、","，","。","？","；","：","‘","“","”", "’"," 【","】","～","！","＠","＃","＄","％","＾","＆","＊","，","．"," ＜", "＞","；","：","＇","＂","［","］","｛","｝","／","＼","《","》","-","_"), array('','','','','','','','','','','','','','','','','','','','','','','','', '','','','','','','','','','','','','','','','','','','','','','','','','','', '','','','','','','','','','','','','',''), $pointer ); }
这个东西有更好的写法，而且全角状态的符号在之前已经过滤了，在此可删除，我懒，抛砖引玉吧
呃，话说我就是把需要匹配的中文符号都列出来的，很笨的方法
[Reply]