如何编写正则表达式，排除句子中目标词组前后的特定词汇关联情况？

Let's refine your regex to handle the exclusion rules more accurately, while covering all your edge cases:

^(?!.*\bno\b\s+\b(?:text|sample text)\b)(?!.*\b(?:text|sample text)\b\s+\bnot\b).+$

Breakdown of the Regex #

• ^ and $: Anchor the match to the start and end of the string, ensuring we check the entire sentence (not just a random substring).
• (?!.*\bno\b\s+\b(?:text|sample text)\b): Negative lookahead to exclude sentences where no (as a standalone whole word) is followed by one or more spaces, then either text or the phrase sample text (both treated as whole-word matches).
• (?!.*\b(?:text|sample text)\b\s+\bnot\b): Second negative lookahead to exclude sentences where text or sample text (whole-word phrases) are followed by one or more spaces, then not (as a standalone whole word).
• .+: Match any non-empty sentence that passes both lookahead exclusion checks.

Verification Against Your Examples #

Let's test this against your sample scenarios:

• ❌ Should NOT match: Sentences with no + sample text (e.g., "I have no sample text here") → The first lookahead catches the invalid pairing and excludes the sentence.
• ❌ Should NOT match: Sentences with text + not (e.g., "This text is not what we want") → The second lookahead triggers the exclusion.
• ❌ Should NOT match: Sentences with no, text, and not (e.g., "No text should not be included") → One of the lookaheads will detect an invalid word pairing and exclude it.
• ✅ Should match: Sentence with only text (e.g., "This is the text we need") → No exclusion conditions are triggered, so it's a valid match.
• ✅ Should match: Generic valid sentence (e.g., "This sentence is a valid match") → No target phrases or exclusion triggers are present, so it's matched.

Notes on Edge Cases #

• The \s+ ensures we account for one or more spaces between trigger words and target phrases (avoids false negatives from accidental word concatenation).
• Using (?:...) for the target phrase group makes it a non-capturing group, keeping the regex efficient and avoiding unnecessary capture clutter.
• If you need to allow for punctuation between words (e.g., "no, text"), you can adjust the regex to include optional punctuation: replace \bno\b\s+ with \bno\b[\s,.]+ to cover commas, periods, or spaces.

内容的提问来源于stack exchange，提问作者Hackerds