去除HTML标签删除HTML示例代码

OStack程序员社区-中国程序员成长平台 › 门户 › 编程› ASP.NET›ASP.NET编程经验

原作者: [db:作者] 来自: [db:来源] 收藏邀请

 
/// <summary> 
/// 去除HTML标签 
/// </summary> 
/// <param name="Htmlstring"></param> 
/// <returns></returns> 
public static string DeleteHTML(string Htmlstring) 
{ 
//删除HTML 
Htmlstring = Regex.Replace(Htmlstring, @"<(.[^>]*)>", "", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"([\r\n])[\s]+", "", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"-->", "", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"<!--.*", "", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(quot|#34);", "\"", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(amp|#38);", "&", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(lt|#60);", "<", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(gt|#62);", ">", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(nbsp|#160);", "", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(iexcl|#161);", "\xa1", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(cent|#162);", "\xa2", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(pound|#163);", "\xa3", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&(copy|#169);", "\xa9", RegexOptions.IgnoreCase); 
Htmlstring = Regex.Replace(Htmlstring, @"&#(\d+);", "", RegexOptions.IgnoreCase); 
Htmlstring = Htmlstring.Replace("<", ""); 
Htmlstring = Htmlstring.Replace(">", ""); 
Htmlstring = Htmlstring.Replace("\r\n", ""); 
return Htmlstring; 
} 