在PHP中使用迭代的嵌套注释

当面对这样的复杂结构时，有时最好创建一个面向对象的解决方案，然后使用对象来创建所需的数组。

例如，基于您的上述内容，我可能会定义以下类：

class Comment{
    protected $id;
    protected $children;
    protected $content;
    public function __construct( $id, $content ){
        $this->id = $id;
        $this->content = $content;
        $this->children = array();
    }
    public function addChild( $child ){
        $this->children[] = $child;
    }
}

现在，我们使用这个对象将您的数据库转移到工作内存中，如下所示：

$workingMemory = array(); //a place to store our objects
$unprocessedRows = array(); //a place to store unprocessed records
// here, add some code to fill $unproccessedRows with your database records
do{
  $row = $unprocessedRows; //transfer unprocessed rows to a working array
  $unprocessedRows = array(); //clear unprocessed rows to receive any rows that we need to process out of order.
  foreach( $row as $record ){
    $id = $record[0]; //assign your database value for comment id here.
    $content = $record[1]; //assign your database value for content here.
    $parentId = $record[2]; //assign your database value for parent id here
    $comment = new Comment( $id, $content );
    //for this example, we will refer to unlinked comments as 
    //having a parentId === null.
    if( $parentId === null ){
         //this is just a comment and does not need to be linked to anything, add it to working memory indexed by it's id.
         $workingMemory[ $id ] = $comment;
    }else if( isset( $workingMemory[ $parentId ] ) ){
         //if we are in this code block, then we processed the parent earlier.
         $parentComment = $workingMemory[ $parentId ];
         $parentComment->addChild( $comment );
         $workingMemory[ $id] = $comment;
    }else{
         //if we are in this code block, the parent has not yet been processed. Store the row for processing again later.
         $unprocessedRows[] = $record;
    }
   }
  }while( count( $unprocessedRows ) > 0 );

一旦所有未处理的行都完成了，现在就可以将注释的表示形式完全存储在变量$workingMemory中，该数组的每个单元格都是一个Comment对象，该对象具有$id、$content，并链接到所有子级$comments。

我们现在可以遍历这个数组，并生成我们想要的任何数据数组或表。我们必须记住，按照我们存储数组的方式，我们可以直接从$workingMemory数组访问任何注释。

如果我用它为网站生成HTML，我会循环遍历workingMemory数组，只处理父注释。然后，每个过程都会遍历子进程。从父母而不是孩子开始，我们可以保证不会两次处理同一条评论。

我会修改我的Comment类以使其更容易：

class Comment{
    protected $id;
    protected $children;
    protected $content;
    protected $isRoot;
    public function __construct( $id, $content ){
        $this->id = $id;
        $this->content = $content;
        $this->children = array();
        $this->isRoot = true;
    }
    public function addChild( $child ){
        $child->isRoot = false;
        $this->children[] = $child;
    }
    public function getChildren(){ return $this->children; }
    public function getId(){ return $this->id; }
    public function getContent(){ return $this->content; }
}

更改后，我可以创建如下HTML：

function outputCommentToHTML( $aComment, $commentLevel = 0 ){
    //I am using commentLevel here to set a special class, which I would use to indent the sub comments.
    echo "<span class'comment {$commentLevel}' id='".($aComment->getId())."'>".($aComment->getContent())."</span>";
    $children = $aComment->getChildren();
    foreach( $children as $child ){
         outputCommentToHTML( $child, $commentLevel + 1 );
    }
}
foreach( $workingMemory as $aComment ){
    if( $aComment->isRoot === true ){
         outputCommentToHTML( $aComment );
    }
}

这将把数据库列转换成您需要的格式。例如，如果我们有以下数据：

comment_id content parent_id
 1          xxx     0
 2          xxx     0
 3          xxx     1
 4          xxx     3
 5          xxx     4
 6          xxx     3
 ...        ...     ...

它将以HTML:输出

  Comment_1
      Comment_3
           Comment_4
                Comment_5
           Comment_6
  Comment_2

这是在函数中递归完成的，该函数在移动到Comment2之前完全处理Comment_1。在进入注释2之前，它还完全处理了Comment_3，这就是注释4、5和6在注释2之前输出的方式。

上面的例子对你有用，但如果这是我的个人项目，我不会混合线性和面向对象的代码，所以我会创建一个代码工厂来将Comments转换为HTML。工厂从源对象生成数据字符串。您可以创建一个充当HTML工厂的Object，以及另一个充当SQL生成器的工厂，通过使用这样的解决方案分层对象，您可以创建完全面向对象的解决方案，这对普通读者来说更容易理解，有时甚至对非编码者来说也更容易理解

//these definition files get hidden and tucked away for future use
//you use include, include_once, require, or require_once to load them
class CommentFactory{
      /**** other Code *****/
      public function createCommentArrayFromDatabaseRecords( $records ){
                  /*** add the data conversion here that we discussed above ****/
                  return $workingMemory;
      }
}
class HTMLFactory{
       public function makeCommentTableFromCommentArray( $array ){
            $htmlString = "";
            foreach( $array as $comment ){
                 if( $comment->isRoot ){
                        $htmlString .= $this->getHTMLStringForComment( $comment );
                 }
            }
            return $htmlString;
       }
       private function getHTMLStringForComment( $comment, $level=0 ){
            /*** turn your comment and all it's children into HTML here (recursively) ****/
            return $html;
       }
}

如果操作得当，它可以清理你的活动代码文件，使其读起来几乎像这样的指令列表：

//let database be a mysqli or other database connection
//let the query function be whatever method works for your database
// of choice.
//let the $fetch_comment_sql variable hold your SQL string to fetch the
//   comments
$records = $database->query( $fetch_comment_sql )
$comFactory = new CommentFactory();
$commentArray = $comFactory->createCommentArrayFromDatabaseRecords( $records );
$htmlFactory = new HTMLFactory();
$htmlResult = $htmlFactory->makeCommentTableFromCommentArray( $commentArray );
echo $htmlResult;