Question

我有一个这样的表，它显示哪个用户评论了哪个父线程。

ParentID  CommentID UserName  CommentDateTime
58         58       Vicky     2016-12-02 11:51:07.270
58         61       Billu     2016-12-02 12:35:40.220
58         62       Rakesh    2016-12-02 12:37:42.133

如果假设对第2行发表评论，则会生成新的荣誉，即63

我想编写一个sql查询，按以下顺序列出行：

ParentID  CommentID UserName  CommentDateTime
58         58       Vicky     2016-12-02 11:51:07.270
58         61       Billu     2016-12-02 12:35:40.220
61         63       Rakesh    2016-12-02 13:37:42.133
58         62       Rakesh    2016-12-02 12:37:42.133

请你帮我写一个逻辑/ SQL吗？

Answer 1

如果我说得对，你想按所需顺序输出记录，不要将它们插入表中。在这里，我们看到了一个帖子的TREE，所以我认为你应该使用recursive CTE。对于每个节点，我们构建一个PATH字符串，然后使用此路径字符串对节点进行排序：

WITH CTE AS
(
   SELECT ParentID, CommentID, UserName, CommentDateTime,
          ParentID as ThreadID,
          CAST(ParentID as varchar(MAX)) as PathStr
   FROM T 
   WHERE CommentID = T.ParentID 

   UNION ALL

   SELECT T.ParentID, T.CommentID, T.UserName, T.CommentDateTime,
          CTE.ThreadID,
          PathStr+'-'
          +CAST(T.CommentID as varchar(MAX)) as PathStr
   FROM T 
   JOIN CTE ON CTE.CommentID = T.ParentID
   WHERE T.CommentID <> T.ParentID 
)

SELECT * FROM CTE ORDER BY ThreadID,PathStr

我已经在示例表中添加了更多节点，因此结果如下：

╔══════════╦═══════════╦══════════╦═════════════════════════╦══════════╦═════════════╗
║ ParentID ║ CommentID ║ UserName ║     CommentDateTime     ║ ThreadID ║   PathStr   ║
╠══════════╬═══════════╬══════════╬═════════════════════════╬══════════╬═════════════╣
║       58 ║        58 ║ Vicky    ║ 2016-12-02 11:51:07.270 ║       58 ║ 58          ║
║       58 ║        61 ║ Billu    ║ 2016-12-02 12:35:40.220 ║       58 ║ 58-61       ║
║       61 ║        63 ║ Rakesh   ║ 2016-12-02 13:37:42.133 ║       58 ║ 58-61-63    ║
║       58 ║        62 ║ Rakesh   ║ 2016-12-02 12:37:42.133 ║       58 ║ 58-62       ║
║      158 ║       158 ║ Vicky    ║ 2016-12-02 11:51:07.270 ║      158 ║ 158         ║
║      158 ║       161 ║ Billu    ║ 2016-12-02 12:35:40.220 ║      158 ║ 158-161     ║
║      161 ║       163 ║ Rakesh   ║ 2016-12-02 13:37:42.133 ║      158 ║ 158-161-163 ║
║      161 ║       164 ║ Rakesh   ║ 2016-12-02 13:37:42.133 ║      158 ║ 158-161-164 ║
║      158 ║       162 ║ Rakesh   ║ 2016-12-02 12:37:42.133 ║      158 ║ 158-162     ║
╚══════════╩═══════════╩══════════╩═════════════════════════╩══════════╩═════════════╝

在SQL中的两行之间插入一行

1 个答案: