Question

我试图从使用HiearchyID的单个表中获取Parent及其左上角和右上角的子项。几个星期以来我一直在喋喋不休，并希望得到任何帮助。

这是一张存储＆＃39;事实的单一表格。对于一个学校辩论小组来说，每个事实都有儿童事实，这些事实可以是“和”。或者反对＆＃39;父母的事实。我希望得到父母和最近的＆＃39; for＆＃39;最近的反对＆＃39;摘要页面的事实。相反，我会与这些组合中的每个组合进行交叉联合。

这是我的表：

{ d '2015-07-05' }

这是我当前的疑问：

factID (key)    nodeID (hierarchyID)    nodeLevel(computed column) text                        side(1=for, 2=against)  timestamp
=======         ======                   =========                 ====                        =====                     =========
1              /1/2/                     2                         "Kirk rules"                1                         08:00
3              /1/2/1/                   3                         "Great actor!"              1                         08:01
5              /1/2/2/                   3                         "Picard is better."         2                         08:02
7              /1/2/3/                   3                         "best captain ever"         1                         08:03
32             /1/2/4/                   3                         "hate his over-acting"      2                         08:04
43             /1/2/5/                   3                         "PriceLine is great."       1                         08:05
44             /1/2/6/                   3                         "Spock was better too."     2                         08:06

这是我想要的：

SELECT P.text AS parentText, P.timeStamp AS parentTimeStamp,  
           L.text AS leftText,  MAX(L.timeStamp) AS leftTimeStamp, 
           R.text AS rightText, MAX(R.timeStamp) AS rightTimeStamp   

FROM app.Facts AS P 
    LEFT OUTER JOIN app.Facts AS L 
            ON (P.nodeID = (L.nodeID).GetAncestor(1))  AND (L.sideID = 1 OR L.sideID IS NULL)

    LEFT OUTER JOIN app.Facts AS R
        ON (P.nodeID = (R.nodeID).GetAncestor(1))  AND (R.sideID = 2 OR R.sideID IS NULL)

WHERE (P.text IS NOT NULL) AND (P.nodeLevel = 2)     

GROUP BY P.text, P.timeStamp, L.text, L.timeStamp, R.text, R.timeStamp
HAVING L.timeStamp = MAX(L.timeStamp) AND R.timeStamp = MAX(R.timeStamp)
ORDER BY P.timeStamp DESC, L.timeStamp DESC, R.timeStamp DESC

这是我得到的：

parentText      parentTimeStamp leftText                 leftTimeStamp  rightText              rightTimeStamp
==========      =============== =====================   =============   =========               ==============
"Kirk rules"    08:00           "PriceLine is great."   08:05          "Spock was better too."  08:06
...
"top fact xx"   11:00           'for' fact xx           11:01          'against' fact xx        11:01

注意：

TOP（1）如果只有一个父母会工作，但我真正的桌子有很多顶级父母。
根事实始终为2级，前两个树级别基本上只是占位符。
我在SQL Server 2014上使用T-SQL。

真的很感激任何提示或解决方案！

Answer 1

我通过添加两列来找到解决方法：

isLatestForArg     BIT  NOT NULL,
isLatestAgainstArg BIT  NOT NULL

然后将我的查询更改为：

SELECT P.text AS parentText, P.timeStamp AS parentTimeStamp,  
       L.text AS leftText,  MAX(L.timeStamp) AS leftTimeStamp, 
       R.text AS rightText, MAX(R.timeStamp) AS rightTimeStamp   

FROM app.Facts AS P 
LEFT OUTER JOIN app.Facts AS L 
        ON (P.nodeID = (L.nodeID).GetAncestor(1))  AND ((L.sideID IS NULL) OR (L.isLatestForArg = 1))

LEFT OUTER JOIN app.Facts AS R
    ON (P.nodeID = (R.nodeID).GetAncestor(1))  AND ((R.sideID IS NULL) OR (R.isLatestAgainstArg = 1))

显然不理想，因为现在每个INSERT都需要更新isLatestFor / Against标志。

查询hierarchyID表的父项和前两个子项

1 个答案: