从选择语句到删除答案

【问题标题】：Going from Select Statement to Delete从选择语句到删除
【发布时间】：2021-02-27 05:46:22
【问题描述】：

我有这个查询，发现有重复记录：

SELECT x.[Constituent Number], x.[First Name], x.[Last or Org Name],x.[Address Line 1], Cnt = Count(*) 
FROM (   
    SELECT SR.[Constituent Number], SR.[First Name], SR.[Last or Org Name], SR.[Address Line 1]
    FROM dbo.['All Salsa Records TL$'] SR

    UNION ALL

    SELECT WD.[Member Number], WD.[First Name], wd.[Last or Org Name], WD.[Address Line 1]
    FROM dbo.['Workling to delete$'] WD
) x
GROUP BY x.[Constituent Number], x.[First Name],x.[Last or Org Name],x.[Address Line 1]
HAVING COUNT(*) > 1

这是在两张桌子之间。我将如何从主表中删除记录以匹配此查询选择的结果？

编辑：我设法用这段代码弄明白了：

DELETE SR
--SELECT SR.[Constituent Number], SR.[First Name], SR.[Last or Org Name]
FROM dbo.['All Salsa Records TL$'] AS SR
LEFT OUTER JOIN dbo.['Workling to delete$'] WD
    ON WD.[Member Number] = SR.[Constituent Number]
    AND WD.[First Name] = SR.[First Name]
    AND WD.[Last or Org Name] = SR.[Last or Org Name] 
    AND WD.[Address Line 1] = SR.[Address Line 1]
    AND WD.[City]  = SR.[City]
    AND WD.[State] = SR.[State]
WHERE SR.[Constituent Number] = WD.[Member Number]
    AND WD.[First Name] = SR.[First Name]
    AND WD.[Last or Org Name] = SR.[Last or Org Name] 
    AND WD.[Address Line 1] = SR.[Address Line 1]
    AND WD.[City]  = SR.[City]
    AND WD.[State] = SR.[State]

对于空值，我只是使用了这个（并进行了相应的更改，因为我有 11 条记录为空成分，26 条记录为空地址）：

DELETE SR
--SELECT SR.[Constituent Number], SR.[First Name], SR.[Last or Org Name]
FROM dbo.['All Salsa Records TL$'] AS SR
LEFT OUTER JOIN dbo.['Workling to delete$'] WD
    ON WD.[Member Number] = SR.[Constituent Number]
    AND WD.[First Name] = SR.[First Name]
    AND WD.[Last or Org Name] = SR.[Last or Org Name] 
    AND WD.[Address Line 1] IS NULL
    AND WD.[City] IS NULL
    AND WD.[State] IS NULL
WHERE SR.[Constituent Number] = WD.[Member Number]
    AND WD.[First Name] = SR.[First Name]
    AND WD.[Last or Org Name] = SR.[Last or Org Name] 
    AND SR.[Address Line 1] IS NULL
    AND SR.[City] IS NULL
    AND SR.[State] IS NULL

【问题讨论】：

标签： sql sql-server sql-delete

【解决方案1】：

如果 ['All Salsa Records TL$'] 是您的主表，而 [Constituent Number] 是您要删除行的列，则：

with cte as(
SELECT x.[Constituent Number], x.[First Name], x.[Last or Org Name],x.[Address Line 1], Cnt = Count(*) 
FROM (   
    SELECT SR.[Constituent Number], SR.[First Name], SR.[Last or Org Name], SR.[Address Line 1]
    FROM dbo.['All Salsa Records TL$'] SR

    UNION ALL

    SELECT WD.[Member Number], WD.[First Name], wd.[Last or Org Name], WD.[Address Line 1]
    FROM dbo.['Workling to delete$'] WD
) x
GROUP BY x.[Constituent Number], x.[First Name],x.[Last or Org Name],x.[Address Line 1]
HAVING COUNT(*) > 1)
delete dbo.['All Salsa Records TL$'] where [Constituent Number]  in (select [Constituent Number] from cte)

【讨论】：

谢谢！我会试试这段代码进行测试！
如果您遇到任何困难，请告诉我。
不用担心 :)。我这样做只是为了测试和学习目的，所以如果我搞砸了也不会造成伤害。这些数据来自我拥有的一个 excel 文件。我运行了您的查询，它删除了 4,565 条记录，而连接删除非常接近。它是 2,471 条记录的选择查询中的 1 条记录。我很好奇 cte 删除的内容是什么推动了更高的记录。不过我可能已经想通了。我对计数 > 1 的组合成分数进行了计数，得到了 2,120。我将 excel 中的计数相加，得到了大约 4,248 的总和。我猜原始数据可能有重复。
我有一个备份，以防万一我需要撤消。
如果“组成编号”有多行，则删除的行数将多于选择，因为选择是基于分组依据进行的。你想删除这些行吗？那么您需要按条件添加所有列。