比如表結構數據如下:
Table:Tree
ID Name ParentId
1 一級 0
2 二級 1
3 三級 2
4 四級 3
SQL SERVER 2005查詢方法:
//上查 with tmpTree as ( select * from Tree where Id=2 union all select p.* from tmpTree inner join Tree p on p.Id=tmpTree.ParentId ) select * from tmpTree //下查 with tmpTree as ( select * from Tree where Id=2 union all select s.* from tmpTree inner join Tree s on s.ParentId=tmpTree.Id ) select * from tmpTree
SQL SERVER 2008及以後版本,還可用如下方法:
增加一列TID,類型設為:hierarchyid(這個是CLR類型,表示層級),且取消ParentId字段,變成如下:(表名為:Tree2)
TId Id Name
0x 1 一級
0x58 2 二級
0x5B40 3 三級
0x5B5E 4 四級
查詢方法:
SELECT *,TId.GetLevel() as [level] FROM Tree2 --獲取所有層級 DECLARE @ParentTree hierarchyid SELECT @ParentTree=TId FROM Tree2 WHERE Id=2 SELECT *,TId.GetLevel()AS [level] FROM Tree2 WHERE TId.IsDescendantOf(@ParentTree)=1 --獲取指定的節點所有下級 DECLARE @ChildTree hierarchyid SELECT @ChildTree=TId FROM Tree2 WHERE Id=3 SELECT *,TId.GetLevel()AS [level] FROM Tree2 WHERE @ChildTree.IsDescendantOf(TId)=1 --獲取指定的節點所有上級
可參見相關文章:http://blog.csdn.net/szstephenzhou/article/details/8277667
ORACLE中的查詢方法:
SELECT * FROM Tree START WITH Id=2 CONNECT BY PRIOR ID=ParentId --下查 SELECT * FROM Tree START WITH Id=2 CONNECT BY ID= PRIOR ParentId --上查
可參見相關文章:http://blog.csdn.net/super_marioli/article/details/6253639
MYSQL 中的查詢方法:
//定義一個依據ID查詢所有父ID為這個指定的ID的字符串列表,以逗號分隔 CREATE DEFINER=`root`@`localhost` FUNCTION `getChildLst`(rootId int,direction int) RETURNS varchar(1000) CHARSET utf8 BEGIN DECLARE sTemp VARCHAR(5000); DECLARE sTempChd VARCHAR(1000); SET sTemp = '$'; IF direction=1 THEN SET sTempChd =cast(rootId as CHAR); ELSEIF direction=2 THEN SELECT cast(ParentId as CHAR) into sTempChd FROM Tree WHERE Id=rootId; END IF; WHILE sTempChd is not null DO SET sTemp = concat(sTemp,',',sTempChd); SELECT group_concat(id) INTO sTempChd FROM Tree where (direction=1 and FIND_IN_SET(ParentId,sTempChd)>0) or (direction=2 and FIND_IN_SET(Id,sTempChd)>0); END WHILE; RETURN sTemp; END //查詢方法: select * from tree where find_in_set(id,getChildLst(1,1));--下查 select * from tree where find_in_set(id,getChildLst(1,2));--上查
補充說明:上面這個方法在下查是沒有問題,但在上查時會出現問題(詳見博問:http://q.cnblogs.com/q/76375/),原因在於我的邏輯寫錯了,存在死循環,現已修正,新的方法如下:
CREATE DEFINER=`root`@`localhost` FUNCTION `getChildLst`(rootId int,direction int) RETURNS varchar(1000) CHARSET utf8 BEGIN DECLARE sTemp VARCHAR(5000); DECLARE sTempChd VARCHAR(1000); SET sTemp = '$'; SET sTempChd =cast(rootId as CHAR); IF direction=1 THEN WHILE sTempChd is not null DO SET sTemp = concat(sTemp,',',sTempChd); SELECT group_concat(id) INTO sTempChd FROM Tree where FIND_IN_SET(ParentId,sTempChd)>0; END WHILE; ELSEIF direction=2 THEN WHILE sTempChd is not null DO SET sTemp = concat(sTemp,',',sTempChd); SELECT group_concat(ParentId) INTO sTempChd FROM Tree where FIND_IN_SET(Id,sTempChd)>0; END WHILE; END IF; RETURN sTemp; END
這樣遞歸查詢就很方便了。
可參見相關文章:http://blog.csdn.net/jackiehome/article/details/6803978
說明:以上知識點均來源於網上,我這裡只是做一個合計總結。