bash - 如何将包含括号的字符串拆分为数组并将子字符串保留在shell脚本中的括号中

Question

我有一个这样的字符串：

a1="a,b,c,(d,e),(f,g)";

如何获得像这样的数组

arr=["a","b","c","d,e","f,g"];

我想用其他字符替换括号之间的逗号并在转换为数组后将其还原

但我不知道如何只替换括号之间的逗号；如何才能做到这一点？

Answer 1

GNU sed 解析器

sed 's/,/\",\"/g;s/(\(.\)\"/\1/g;s/\"\(.\))/\1/g;s/^\w\+=\"/arr=[\"/;s/;/];/'

Answer 2

写一个解析器！:D

我不知道如何在 bash 中执行此操作，但我可以向您展示如何在 PHP 中执行此操作（应该可以转换为其他语言）。

$str = "a,b,c,(d,e),(f,g)";
$out = array();
$current_token = "";
$open_brackets = 0;
$length = strlen($str)
for ($i = 0; $i < $length; $i += 1) {
    $chr = $str[$i];
    if ($chr === "(") {
        $open_brackets += 1;
    } else if ($chr === ")") {
        $open_brackets -= 1;
    } else if ($open_brackets === 0 && $chr === ",") {
        $out[] = $current_token; // push token value to out
        $current_token = "";
    } else {
        $current_token .= $chr;
    }
}
if (strlen($current_token) > 0) {
    $out[] = $current_token; // dont forget the last one
}
var_dump($out); // ["a","b","c","d,e","f,g"]

未经测试，但这是大纲。跟踪括号的数量，只有当括号匹配时，才应将,其解释为分隔符。

Answer 3

尝试使用正则表达式解析字符串的bash脚本。这对我来说很尴尬，但似乎有效：

#!/usr/bin/env bash

unset arr
a1="a,b,c,xxx(d,e),sdf(f,g)"

## The regular expression does an alternation between
## a pair of parens followed by an optional comma "\([^\)]+\)(,?)"
## or any characters followed by a comma or end of line "[^,]+(,|$)"
## After that I save all the rest of the string to match it in
## following iterations.
while [[ $a1 =~ ([^\(,]*\([^\)]+\)(,?)|[^,]+(,|$))(.*) ]]; do

    ## BASH_REMATCH keeps grouped expressions. The first one
    ## has the data extracted between commas. This removes the
    ## trailing one.
    elem="${BASH_REMATCH[1]%,}"

    ## Remove opening paren, if exists one.
    elem="${elem/\(/}"

    ## Remove trailing paren, if exists one.
    elem="${elem%)}"

    ## Add element to an array.
    arr+=("$elem")

    ## Use the string left (fourth grouped expression in
    ## the regex) to continue matching elements.
    a1="${BASH_REMATCH[4]}"
done

printf "%s\n" "${arr[@]}"

像这样运行它：

bash script.sh

它产生：

a
b
c
xxxd,e
sdff,g