php - 如何编写一个拆分字符串但忽略分隔符的转义序列的正则表达式？

Question

我希望使用 PHPmb_split来拆分以冒号分隔的字符串，但忽略转义的冒号（即\:）。

例如：

Part A is here:Part B\: is over here:Part C\ is here:Part \\:D is here

需要拆分成：

1. Part A is here
2. Part B: is over here
3. Part C\ is here
4. Part \:D is here

我真的不知道该怎么做。有什么建议吗？

Answer 1

You can try splitting on:

(?<!\\):

(?<!\\) is a negative lookbehind and will make : not match if it is preceeded by \.

[Note I never tried mb_split, but this should work with preg_split]

Answer 2

这是我供搜索者参考的完整解决方案。

<?php
if (!function_exists('mb_str_replace')) {
    function mb_str_replace ($needle, $replacement, $haystack) {
        // Courtesy of marc at ermshaus dot org: http://www.php.net/manual/en/ref.mbstring.php#86120
        $needleLength = mb_strlen($needle);
        $replacementLength = mb_strlen($replacement);
        $pos = mb_strpos($haystack, $needle);
        while ($pos !== false) {
            $haystack = mb_substr($haystack, 0, $pos).$replacement.mb_substr($haystack, $pos+$needleLength);
            $pos = mb_strpos($haystack, $needle, $pos+$replacementLength);
        }
        return $haystack;
    }
}

$str = <<<EOD
Part A is here:Part B\: is over here:Part C\ is here:Part \\\:D is here
EOD;
    /* Why did I have to use "\\\:D" in a heredoc? No idea.
       I thought "\\:D" would be sufficient, but for some
       reason the backslash was parsed anyways. Which kinda
       makes heredocs useless to me. */

echo "$str<br>", PHP_EOL;

$arr = mb_split('(?<!\\\):', $str);
    // Courtesy of Jerry: http://stackoverflow.com/users/1578604/jerry
foreach ($arr as &$val) {
    $val = mb_str_replace('\:', ':', $val);
}
unset($val);

echo '<pre>', print_r($arr, true), '</pre>';
?>

输出：

A部分在这里：B部分\：在这里结束：C部分在这里：部分\\：D在这里
大批
(
    [0] => A 部分在这里
    [1] => B 部分：在这里
    [2] => C 部分在这里
    [3] => 部分 \:D 在这里
)