既然您说传统的 html 解析方法似乎不适用于您的用例，您是否考虑过为这个特定用例编写手动解析器？
我写了一个简短但有效的示例，说明您可能需要考虑的内容。
但是要记住几件事，这个实现写得很快，没有错误处理，假设目标字符串可以容纳在内存中，并且缺少关键的边缘情况。如果您想考虑这个解决方案，您应该投入时间来填补空白。
此解决方案仅解析整个文档，忽略并写入非目标字符，当目标字符串被识别时，改为写入替代字符串。
这可能不是最好的解决方案，我鼓励您寻找提供更多功能的现成 HTML 解析器。
public static void RemoveTargetString(TextReader Reader, TextWriter Writer, string TargetString, char[] CharactersToIgnore, string ReplacementString)
{
HashSet<char> IgnoreCases = CharactersToIgnore.ToHashSet();
// our buffer need only be the size of the target string
char[] buffer = new char[TargetString.Length];
int currentIndex = 0;
while (Reader.Peek() > -1)
{
// read one character to the end of the buffer marked by index
if (Reader.Read(buffer, currentIndex, 1) != 0)
{
// get the last char in the buffer
ref char firstChar = ref buffer[currentIndex];
// if the char is on the ignore list blindly write it and continue
// dont change index so we overwrite the char in the last spot of the buffer
if (IgnoreCases.Contains(firstChar))
{
// write the char and ignore
Writer.Write(firstChar);
continue;
}
// check to see if the char is in the right order as the target string
if (firstChar == TargetString[currentIndex])
{
// if it is don't write the buffer, increment index so we keep the char without back tracking
currentIndex++;
// if we have found the entire string dump the buffer, write the replacement string
if (currentIndex == TargetString.Length)
{
// write replacement string instead
Writer.Write(ReplacementString);
// reset index so we overwrite the buffer
currentIndex = 0;
}
}
else
{
// check to see if the target string is within something that starts with a partial piece of the target string
// we should not implicitly assume the character we fail at isn't the start of the target as well
// if it is we should avoid writing it
if (firstChar == TargetString[0])
{
Writer.Write(buffer, 0, currentIndex);
buffer[0] = buffer[currentIndex];
// reset index and start searching for start of target
currentIndex = 1;
}
else
{
// since the char at the last position of the buffer wasn't
// either the start or within the target string
// write the buffer from 0 - last index
Writer.Write(buffer, 0, currentIndex + 1);
// reset index and start searching for start of target
currentIndex = 0;
}
}
}
}
// if for some reason the target string is at the end, but was not complete, we should write the characters in the buffer to the target
if (currentIndex > 0)
{
Writer.Write(buffer, 0, currentIndex);
}
}
char[] IgnoreCharacters = new char[] { '\n', '\r', ' ' };
string target = "<td style=\"text - align: right; \">\n\r<a rel=\"nofollow\" target=\"_blank\" href=\"https://www.monstermmorpg.com\"\n\r style=\"color: rgb(6, 69, 173);\n\r text-decoration-line: none;\n\r background: none;\"\n\r title=\"Calyrex (Pokémon)\"><span style=\"color: rgb(0, 0, 0);\">←</span></a></td>";
StringReader reader = new($"<tr>\n\r<td colspan=\"2\" style=\"background - image: initial; background - position: initial; background - size: initial; background - repeat: initial; background - attachment: initial; background - origin: initial; background - clip: initial; border - top - left - radius: 3px; border - top - right - radius: 3px; \">\n\r<b>{target}</b>\n\r</td>\n\r</tr>");
foreach (char item in IgnoreCharacters)
{
target = target.Replace(item.ToString(), "");
}
StringWriter writer = new();
RemoveTargetString(reader, writer, target, IgnoreCharacters, "AAA");
Console.WriteLine(writer.ToString());
如果您不熟悉TextReader 或TextWriter，这些是StreamReader 和StreamWriter 等常见IO 功能的基类。您可以使用它来简化文件中的行查找信息，如下所示：
char[] IgnoreCharacters = new char[] { '\n', '\r', ' ' };
string target = "Hello World";
string replacement = "Hello Globe";
using StreamReader reader = new("Test.txt");
using StreamWriter writer = new("Output.txt");
RemoveTargetString(reader, writer, target, IgnoreCharacters, replacement);
编辑：
修复了如果目标正在被识别但失败，则单个字符未写入输出流导致有损转录的问题。为常见的边缘案例创建了测试用例。
对于那些对该解决方案的性能感兴趣的人来说，处理一个 1,278,518,583 字节 (1.19GB) 的文本文件需要大约 35 秒，并使用 9 MB 内存。如果需要额外的性能，请考虑将 IgnoreCases.Contains(firstChar) 替换为 Char.IsWhiteSpace(firstChar)，这会快约 33%。
static char[] IgnoreCharacters = new char[] { '\n', '\r', ' ', '\t' };
[Theory]
[InlineData("1234", "<div>1234</div>", "<div></div>")]
[InlineData("1234", "<div>1\n\r2\t3\n4\r</div>", "<div>\n\r\t\n\r</div>")]
[InlineData("1234", "\n\r\t1\n\r\t2\r\n\t3\n\t\r4", "\n\r\t\n\r\t\r\n\t\n\t\r")]
[InlineData("1234", "1 2 3 4", " ")]
[InlineData("1234", " \n\r\t1 \n\r\t2 \n\r\t3 \n\r\t4 \n\r\t", " \n\r\t \n\r\t \n\r\t \n\r\t \n\r\t")]
[InlineData("1234", "123412341234", "")]
[InlineData("1234", "4321", "4321")]
[InlineData("1234", "Hello", "Hello")]
[InlineData("1234", "", "")]
[InlineData("1234", "1/2/3/4", "1/2/3/4")]
[InlineData("1", "1111", "")]
[InlineData("1", "12131415", "2345")]
[InlineData("Abcde", "AbcdAbcde", "Abcd")]
[InlineData("Abcde", "AbcdAbcdeAbcd", "AbcdAbcd")]
[InlineData("12345", "121231234123451234123121", "1212312341234123121")]
public void CommonEdgeCases(string Target, string Input, string Expected)
{
foreach (char item in IgnoreCharacters)
{
Target = Target.Replace(item.ToString(), "");
}
StringReader reader = new(Input);
StringWriter writer = new();
RemoveTargetString(reader, writer, Target, IgnoreCharacters, string.Empty);
Assert.Equal(Expected, writer.ToString());
}