Python正则：如何精准匹配前后非字母的单位（Pa/K/degC/℃）

Hey there! Let's break down how to fix your regex issue so you can match units like Pa, K, degC, and ℃ without grabbing extra characters, and ensure they aren't part of longer words.

为什么你的尝试没生效？ #

First off, the mistake in your second regex is mixing up the syntax for lookbehind (checking what's before the unit) and lookahead (checking what's after). You used (?![A-Za-z]) (negative lookahead) for both the front and back, but to validate the characters before the unit, you need a negative lookbehind: (?<!...). The (?!) syntax only checks what comes after the match, so your regex wasn't actually verifying the preceding characters correctly.

正确的零宽断言正则 #

We need to use zero-width assertions here—these check for conditions without consuming any characters, so you'll only get the unit itself as the match, not surrounding spaces or symbols.

Here's how to write it for each unit, or combined into one regex:

单个单位的正则

• For degC (case-insensitive):

  (?<!\p{L})[Dd][Ee][Gg][Cc](?!\p{L})
  

  Or if your regex engine supports case-insensitive flags (like /i in JavaScript/PCRE), simplify it to:

  (?<!\p{L})degC(?!\p{L})
  

• For Pa:

  (?<!\p{L})Pa(?!\p{L})
  

• For K:

  (?<!\p{L})K(?!\p{L})
  

• For ℃ (or its Unicode escape \u2103):

  (?<!\p{L})℃(?!\p{L})
  

合并所有单位的正则

To match any of these units in one go, use the | (alternation) operator:

(?<!\p{L})(Pa|K|[Dd][Ee][Gg][Cc]|℃)(?!\p{L})

Or with case-insensitive flag:

(?<!\p{L})(Pa|K|degC|℃)(?!\p{L})

关键细节解释 #

• (?<!\p{L}): Negative lookbehind that ensures there is no Unicode letter (\p{L} matches any letter from any language, more comprehensive than [A-Za-z]) immediately before the unit.
• (?!\p{L}): Negative lookahead that ensures there is no Unicode letter immediately after the unit.
• Using \p{L} instead of [A-Za-z] is better because it covers accented letters, non-Latin scripts, etc., preventing false matches in multilingual text. If you only need to exclude English letters, replace \p{L} with [A-Za-z].

测试示例 #

If you run this regex against the text:

"The temperature is 25 degC, pressure is 101325 Pa, and 0 K is absolute zero. Don't match degCelsius or Pawn."

It will correctly match degC, Pa, and K—and ignore the degC in degCelsius and Pa in Pawn, since those are surrounded by letters.

内容的提问来源于stack exchange，提问作者AlexT