Bash脚本在正则表达式上不匹配 [英] Bash script wont match on regular expression

问题描述

我有以下bash脚本，应该生成输出TEST

I have the following bash script which should be producing the output TEST

#!/bin/bash

test="TEST:THING - OBJECT_X"
if [[ $test =~ ^([a-zA-Z0-9]+)\:([a-zA-Z0-9]+)[A-Z\s\-_]+$ ]]; then
      echo ${BASH_REMATCH[1]}
fi

在我的正则表达式测试器中，正则表达式似乎在第一组和第二组上匹配并捕获:

In my regex tester the regular expression seems to be matching and capturing on the first and second groups:

https://regex101.com/r/kR1jM7/1

您知道是什么原因造成的吗?

Any idea whats causing this?

推荐答案

\s是在ERE内部没有意义的PCRE构造.使用[:space:]代替.另外，不要将破折号转义为\-，而是将-移到字符集定义的最后.

\s is a PCRE construct not meaningful inside of ERE. Use [:space:] instead. Also, instead of escaping the dash as \-, move the - to the very end of the character set definition.

以下作品:

[[ $test =~ ^([a-zA-Z0-9]+):([a-zA-Z0-9]+)[A-Z[:space:]_-]+$ ]]

也就是说，为了与更大范围的bash版本兼容，请将正则表达式移至变量中:

That said, for compatibility with a wider range of bash releases, move the regex into a variable:

re='^([a-zA-Z0-9]+):([a-zA-Z0-9]+)[A-Z[:space:]_-]+$'
[[ $test =~ $re ]]

---

要更积极地使用POSIX字符类(从而使您的代码更有可能在各种语言和语言环境中正常工作)，请考虑:

---

To use POSIX character classes more aggressively (and thus make your code more likely to work correctly across languages and locales), also consider:

re='^([[:alnum:]]+):([[:alnum:]]+)[[:upper:][:space:]_-]+$'

这篇关于Bash脚本在正则表达式上不匹配的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！