我设法通过以下方式完成了我想做的事情（ 不使用Regex ）：
- 解析字符串的每个字符
- 删除所有
<a>
标签（将它们复制到临时数组并在字符串上保留占位符） str_replace
新字符串以替换所有关键字- 通过原始
<a>
标签重新填充占位符
这是我使用的代码，以防其他人需要它：
$str = <<<STRAMoses supposes his toeses are roses,but <a href="original-moses1.html">Moses</a> supposes erroneously;for nobody's toeses are posies of roses,as Moses supposes his toeses to be.Ganda <span ><a href="original-moses2.html" target="_blank">Moses</a></span>!STRA;$arr1 = str_split($str);$arr_links = array();$phrase_holder = '';$current_a = 0;$goto_arr_links = false;$close_a = false;foreach($arr1 as $k => $v){ if ($close_a == true) { if ($v == '>') { $close_a = false; } continue; } if ($goto_arr_links == true) { $arr_links[$current_a] .= $v; } if ($v == '<' && $arr1[$k+1] == 'a') { // keep collecting every char until </a> $arr_links[$current_a] .= $v; $goto_arr_links = true; } elseif ($v == '<' && $arr1[$k+1] == '/' && $arr1[$k+2] == 'a' && $arr1[$k+3] == '>' ) { $arr_links[$current_a] .= "/a>"; $goto_arr_links = false; $close_a = true; $phrase_holder .= "{%$current_a%}"; $current_a++; } elseif ($goto_arr_links == false) { $phrase_holder .= $v; }}echo "links Array:n";print_r($arr_links);echo "nnnPhrase Holder:n";echo $phrase_holder;echo "nnn(pre) Final Phrase (with my keyword replaced):n";$final_phrase = str_replace("Moses", "<a href="novo-mega-link.php">Moses</a>", $phrase_holder);echo $final_phrase;echo "nnnFinal Phrase:n";foreach($arr_links as $k => $v){ $final_phrase = str_replace("{%$k%}", $v, $final_phrase);}echo $final_phrase;输出：
链接数组：
Array( [0] => <a href="original-moses1.html">Moses</a> [1] => <a href="original-moses2.html" target="_blank">Moses</a>)
词组：
Moses supposes his toeses are roses,but {%0%} supposes erroneously;for nobody's toeses are posies of roses,as Moses supposes his toeses to be.Ganda <span >{%1%}</span>!（上）最终词组（替换为我的关键字）：
<a href="novo-mega-link.php">Moses</a> supposes his toeses are roses,but {%0%} supposes erroneously;for nobody's toeses are posies of roses,as <a href="novo-mega-link.php">Moses</a> supposes his toeses to be.Ganda <span >{%1%}</span>!最终词组：
<a href="novo-mega-link.php">Moses</a> supposes his toeses are roses,but <a href="original-moses1.html">Moses</a> supposes erroneously;for nobody's toeses are posies of roses,as <a href="novo-mega-link.php">Moses</a> supposes his toeses to be.Ganda <span ><a href="original-moses2.html" target="_blank">Moses</a></span>!
