JavaScript实现大字符串与字符串数组的3词及以上连续匹配问题

我有一个包含数千个单词的大字符串，需要与数组中同样为大字符串的所有元素进行对比，找出所有3个及以上连续匹配的单词。用正则表达式实现后得到的却是空匹配数组。

示例（简化文本） #

let textToCompare = "Hello there how are you doing with your life";

let textsToCompareWith= [
  { id:1, text:"Hope you are doing good with your life" },
  { id:2, text:"what are you doing with your life. hello there how are you" },
  { id:3, text:"hello there mate" }
];

预期输出 #

[
  {id:1, matchedText:["with your life"]}, 
  {id:2, matchedText:["are you doing with your life","hello there how are you"]},
  {id:3, matchedText:[]}
];

当前输出 #

[
  {id:1, matchedText:[]}, 
  {id:2, matchedText:[]},
  {id:3, matchedText:[]}
];

我的代码 #

let regex = new RegExp("\\b" + textToCompare.split(" ").join("\\b.*\\b") + "\\b", "gi");

let output = textsToCompareWith.map(textObj => {
  // Match against each element in the array
  let matchedText = textObj?.text.match(regex);
  console.log(matchedText);
  return {
    id: textObj.id,
    matchedText: matchedText ? matchedText : [] // Return an empty array if no match is found
  };
});
                                  
console.log(output);

你的正则表达式逻辑完全错误：它会要求匹配整个textToCompare的单词顺序，中间允许插入任意字符，而不是从目标文本中找出原字符串里连续3个及以上单词的片段。要实现需求，需要换思路：

核心思路 #

1. 先从原字符串中生成所有长度≥3的连续单词组合（比如"Hello there how"、"are you doing with your life"等）
2. 对每个组合生成正则，在目标文本中匹配完整单词序列，最后去重整理结果

修正后的代码 #

const textToCompare = "Hello there how are you doing with your life";
const sourceWords = textToCompare.split(" ");

// 生成所有长度≥3的连续单词组合
const candidatePhrases = [];
for (let i = 0; i <= sourceWords.length - 3; i++) {
  for (let j = i + 2; j < sourceWords.length; j++) {
    candidatePhrases.push(sourceWords.slice(i, j + 1).join(" "));
  }
}

// 去重，减少重复匹配次数
const uniquePhrases = [...new Set(candidatePhrases)];

const textsToCompareWith= [
  { id:1, text:"Hope you are doing good with your life" },
  { id:2, text:"what are you doing with your life. hello there how are you" },
  { id:3, text:"hello there mate" }
];

const output = textsToCompareWith.map(textObj => {
  const matched = [];
  uniquePhrases.forEach(phrase => {
    // 用单词边界确保完整单词匹配，不区分大小写
    const regex = new RegExp(`\\b${phrase}\\b`, "gi");
    const matches = textObj.text.match(regex);
    if (matches) {
      // 避免同一文本重复添加相同匹配项
      matches.forEach(match => {
        const lowerMatch = match.toLowerCase();
        if (!matched.some(item => item.toLowerCase() === lowerMatch)) {
          matched.push(match);
        }
      });
    }
  });
  // 按长度降序排列，优先保留更长的匹配片段（可选）
  matched.sort((a, b) => b.split(" ").length - a.split(" ").length);
  return {
    id: textObj.id,
    matchedText: matched
  };
});

console.log(output);

代码说明 #

• 生成候选片段：遍历原字符串的单词数组，生成所有长度从3到总单词数的连续组合，覆盖所有可能的匹配项
• 去重处理：避免重复生成相同短语，减少不必要的正则匹配
• 正则匹配：用\b包裹短语，确保匹配完整单词，不会出现you匹配your的情况
• 结果去重：同一文本中可能多次出现相同匹配片段，去重后返回更简洁的结果

性能优化提示 #

如果原字符串包含上万级别的单词，预生成所有候选片段会占用大量内存。此时可以改用滑动窗口的方式，直接在目标文本中拆分单词，然后与原字符串的单词数组做连续匹配对比，避免预生成所有组合。

内容的提问来源于stack exchange，提问作者BSA