Java如何修改字符串仅保留末尾带标点符号的句子

• 你当前使用的 String.replace() 仅支持固定字符串的精确匹配替换，无法实现这种需要规则判断的文本过滤需求。要实现仅保留带结束标点（./?/!）句子的需求，更推荐用正则匹配提取所有合法句子后拼接的方案，稳定性更高。

推荐实现代码 #

import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class FilterValidSentences {
    public static void main(String[] args) {
        String text = "This is some text. I want to keep certain portions of it. I only want to keep sentences with punctuation. This is not an example of a sentence that should be kept This is an example of text that should be kept. How would I do this?";
        // 正则规则：匹配任意非结束标点的字符序列，后跟一个结束标点，兼容标点后的空格/换行
        Pattern validSentencePattern = Pattern.compile("[^.!?]+[.!?]\\s*");
        Matcher matcher = validSentencePattern.matcher(text);
        StringBuilder resultBuilder = new StringBuilder();

        while (matcher.find()) {
            resultBuilder.append(matcher.group());
        }

        // 最终结果，可按需调用trim()去除首尾多余空白
        String filteredText = resultBuilder.toString().trim();
        System.out.println(filteredText);
    }
}

输出结果 #

执行上述代码后输出内容如下，已自动剔除无结束标点的无效句子：

This is some text. I want to keep certain portions of it. I only want to keep sentences with punctuation. This is an example of text that should be kept. How would I do this?

可选：直接使用replaceAll实现 #

如果希望用replaceAll方法一行实现，可使用如下写法（稳定性略低于提取方案）：

text = text.replaceAll("(?<=\\s|^)[^.!?]+(?=\\s[^.!?])", "").replaceAll("\\s{2,}", " ").trim();

规则扩展说明 #

如果需要适配中文标点（。/？/！）等其他结束符，只需把正则中的[.!?]替换为[.!?。？！]即可。

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