我真的试着想出来了，我得出的结论是，真的没有其他天生的方法可以做到这一点。您的方法可能是另一条路线，但目前来看，运行起来相当昂贵。

我得出的结论是：

在一个与帖子类型和分类相关联的查询中获取所有帖子。为了流式处理查询，我们将只查询帖子ID。这将大大提高性能，因为它可以节省高达99.9%的db查询和运行这些查询所需的时间。

我们将循环查看这些帖子ID并使用get_the_terms() 获取与帖子相关的术语。应注意的是get_the_terms() 缓存，所以这也是非常精简的。

然后，我们将循环浏览帖子术语，并将它们存储在一个数组中，然后将它们与相邻帖子的术语进行比较。前三个帖子ID将保存在一个数组中。这将是最近的3篇帖子，每一篇都来自一个独特的类别。为了加速循环，我们将使用break 一旦我们找到了我们想要的

我们还将实现一些系统，以避免在每次加载页面时都运行该系统

好的，让我们来解决这个问题并编写代码。在我们开始之前，请注意几点。

重要提示：

• 该代码完全未经测试。这可能是马车。请确保在本地测试并启用调试

• 至少需要PHP 5.4。

  根据需要修改和滥用代码

  我们将使用一个函数，稍后可以在适当的位置调用该函数

  /**
   * Function get_unique_term_recent_posts
   *
   * Get most recent posts but each post should be from a
   * unique term
   *
   * @see http://wordpress.stackexchange.com/q/206100/31545
   * @param (string) $post type Post type to get posts from
   * @param (string) $taxonoy Taxonomy to get posts from
   * @param (int|string) $number Amount of posts to get
   * @return $unique_ids
   *
   */
  function get_unique_term_recent_posts( $post_type = \'post\', $taxonomy = \'category\', $number = 3 )
  {
      // Make sure that no empty values are passed, if so, return false
      if (    !$post_type 
           || !$taxonomy
           || !$number
      ) 
          return false;
  
      // Validate and sanitize input values
      $post_type = filter_var( $post_type, FILTER_SANITIZE_STRING );
      $taxonomy  = filter_var( $taxonomy,  FILTER_SANITIZE_STRING );
      $number    = filter_var( $number,    FILTER_VALIDATE_INT    );
  
      // Make sure that our taxonomy exist to avoid WP_Error objects when querying post terms, return false
      if ( !taxonomy_exists( $taxonomy ) )
          return false;
  
      // Save everything in a transient to avoid this being run on every page load
      if ( false === ( $unique_ids = get_transient( \'posts_list_\' . md5( $taxonomy . $post_type . $number ) ) ) ) {
  
          // Run our query to get the posts
          $args = [
              \'post_type\'      => $post_type,
              \'posts_per_page\' => -1,
              \'fields\'         => \'ids\' // only get post ids, all we need here
          ];
          $q = get_posts( $args );
  
          // Make sure we have posts before we continue, if not, return false
          if ( !$q )
              return false;
  
          // Lets update the object term cache
          update_object_term_cache( $q, $post_type );
  
          // Set up the varaible which will hold the post ID\'s
          $unique_ids = [];
          // Set up a variable which will hold term ID for comparion
          $term_id_array = [];
          // Set up a helper variable which will hold temp term ID\'s for comparion
          $term_helper_array = [];
          // Start a counter to count posts with terms from $taxonomy
          $counter = 0;
  
          // Loop through the posts, get the post terms, loop through them and build an array of post ID\'s
          foreach ( $q as $post_id ) { 
  
              // Break the foreach loop if the amount of ids in $unique_ids == $number
              if ( count( $unique_ids ) == $number )
                  break;
  
              $terms = get_object_term_cache( $post_id, $taxonomy );
  
              // If post does not have terms, skip post and continue
              if ( !$terms ) 
                  continue;
  
              /**
               * If $term_id_array is empty, this means that this is the first post/newest
               * post that will have a unique term, lets save the post ID
               */
              if (    !$term_id_array
                   && $counter == 0
              )
                  $unique_ids[] = $post_id;
  
              // We do have post terms, loop through them
              foreach ( $terms as $term_key=>$term ) {
  
                  // Add all term ID\'s in array. Skip comparison on first post, just save the terms
                  if ( $counter == 0 ) {
                      $term_id_array[] = $term->term_id;
                  } else { 
  
                      if ( in_array( $term->term_id, $term_id_array ) ) {
  
                          // Reset the $term_helper_array back to an empty array
                          $term_helper_array = [];
  
                          // Break the loop if the condition evaluate to true
                          break;
                      } 
                          // Term ID not found, update the $term_helper_array with the current term ID
                          $term_helper_array[] = $term->term_id;
  
                  } //endif $counter == 0
  
              } // endforeach $terms
  
              /**
               * If our helper array, $term_helper_array have terms, we should move them to the $term_id_array array
               * and then reset the $term_helper_array back to an empty array. 
               *
               * In short, if $term_helper_array have terms, it means that the specific post has unique terms
               * compare to the previous post being save in $unique_ids
               */
              if ( $term_helper_array ) {
                  // If no term ID mathes an ID in the $term_id_array, save the post ID
                  $unique_ids[] = $post_id;
                  /**
                   * Merge the $term_id_array and $term_helper_array. This will hold
                   * only terms of the post that are stored in the $unique_ids array
                   */
                  $term_id_array = array_merge( $term_id_array, $term_helper_array );
  
                  // Reset the $term_helper_array back to an empty array
                  $term_helper_array = [];
              }
  
              // Update our counter
              $counter++;
  
          } // endforeach $q   
  
      set_transient( \'posts_list_\' . md5( $taxonomy . $post_type . $number ), $unique_ids, 30 * DAY_IN_SECONDS );
      }
      /**
       * We can now return $unique_ids which should hold at most 3 post ID\'s, 
       * each having a unique term    
       */
      return $unique_ids;
  }   
  

  此功能应该可以完美地用于分配了多个术语的帖子（通常情况下是这样的）。如果一个帖子a分配了术语a和术语B，而在帖子B分配了术语B和术语C，那么帖子B将不会显示在返回的ID数组中，因为您希望有两个帖子分配了唯一的术语。

  您现在可以按如下方式使用该功能：

  $post_type = \'webcomic1\';
  $taxonomy = \'webcomic1_storyline\';
  $post_ids = get_unique_term_recent_posts( $post_type, $taxonomy );
  if ( $post_ids ) {
      $args = [
          \'post__in\' => $post_ids,
          \'post_type\' => $post_type,
          \'posts_per_page\' => 3
      ];
      $q = new WP_Query( $args );
  
      if ( $q->have_posts() ) {
          while ( $q->have_posts() ) {
          $q->the_post();
  
              // Your template tags and mark up
  
          } // endwhile
          wp_reset_postdata();
      } // endif have_posts()
  } // endif $post_ids
  

  我已将瞬态设置为在30天内过期，您可以根据需要将其设置为。但要了解更多要点控制，您可能希望在发布、删除、更新或取消删除新帖子时删除此瞬态。为此，我们将在transition_post_status 钩火。他的钩子在所有先前命名的条件下开火。

  您可以执行以下操作

  add_action( \'transition_post_status\', function ()
  {
      global $wpdb;
      $wpdb->query( "DELETE FROM $wpdb->options WHERE `option_name` LIKE (\'_transient%_posts_list_%\')" );
      $wpdb->query( "DELETE FROM $wpdb->options WHERE `option_name` LIKE (\'_transient_timeout%_posts_list_%\')" );
  });
  

  编辑该代码现在已经过测试，并按预期运行，修复了几个bug。

  只是关于函数的性能

  在0.27秒内无瞬态->2db查询。

  在0.007秒内完成瞬态->2db查询。

  正如您所看到的，db调用保持不变，临时调用的唯一原因是为了节省内存，因为函数在没有临时调用的情况下会使用一点内存来运行，而且，使用的PHP函数也会花费大量时间来执行。因此，本质上，瞬态是用来节省时间和内存的，这在非常大的站点上非常有用