如何在同一正则表达式中实现分隔符的含/不含区分并适配多行文本？

你现在用C#的Regex.Split处理字符串，代码如下：

string StringX = TextBoxY.Text;

string[] splicharX = Regex.Split(StringX, @"\s(today.+|but|)\s");

场景1 #

输入文本：

Today they are really tired but tomorow they will go back to work.

你想要的输出是：

Today they are really tired

场景2 #

如果把正则改成@"\s(today.+|but|.+back)\s"，同样的输入文本，你想要的输出是：

Today they are really tired 
tomorow they will go back

核心需求：要让today.+、.+back这类匹配的内容留在分割结果里，but这类直接排除，还要能适配多行文本。怎么写单行正则实现这个？

要搞定这个，关键是利用正则的捕获组和非捕获组的区别，再加上单行模式适配多行，具体思路和代码如下：

正则写法 #

直接用这个正则就行：

@"(?s)\s+(?:but\s+)|(\s*(?:today.*?|.*?back)\s*)"

完整代码示例 #

string input = @"Today they are really tired but tomorow they will go back to work.
Another line today starts here and ends back here.";

// 分割+过滤空内容+去首尾空格
var result = Regex.Split(input, @"(?s)\s+(?:but\s+)|(\s*(?:today.*?|.*?back)\s*)")
    .Where(s => !string.IsNullOrWhiteSpace(s))
    .Select(s => s.Trim())
    .ToArray();

// 打印结果
foreach (var piece in result)
{
    Console.WriteLine(piece);
}

正则逻辑解释 #

• (?s)：开启单行模式，让.能匹配换行符，直接适配多行文本；
• \s+(?:but\s+)：用非捕获组(?:...)包裹but，意思是“匹配到这个就分割，但别把它放进结果里”，\s+处理前后的空格；
• (\s*(?:today.*?|.*?back)\s*)：用捕获组()包裹你想要保留的内容——要么是从today开始的片段，要么是到back结束的片段，.*?是非贪婪匹配，避免吃掉后面的目标内容；
• 最后用Where和Select清理掉分割出来的空字符串和多余空格，得到干净的结果。

实际效果 #

对于多行输入：

Today they are really tired but tomorow they will go back to work.
Another line today starts here and ends back here.

输出是：

Today they are really tired
tomorow they will go back
today starts here and ends back

内容的提问来源于stack exchange，提问作者user20465780