sqldf 仅返回一行，与 SQL 中使用的查询相同

Question

由于某种原因，当涉及到 R 时，我只返回一行，而在 SQL Server 中，我返回了正确的行数。 SQLDF：

CustomerCodingChangesT <- sqldf("
        SELECT c.CustID as ID 
             , c.ReverseSupplier as Supplier
             , c.ReverseCustomerCode as Code
             , c.Name
             , c.Address
             , c.[From PostCode]
             , c.[From Outlet]
             , c.[From OutletName] 
             , o.FullAddress AS [From Address]
             , c.[To PostCode]
             , c.[To Outlet]
             , c.[To OutletName]
             , o1.FullAddress AS [To Address]
             , MAX(CAST(c.TotalUnits AS VARCHAR)) AS [Total Units]
             , '$'+MAX(CAST(c.TotalValue AS VARCHAR)) AS [Total Value]
             , '' AS Checked
             , c.CustRecActive as Active

        FROM CustomerCorrectionSummaryT AS c
        LEFT JOIN OutletMasterT AS o 
             ON  c.[From PostCode] = o.Postcode 
             AND c.[From Outlet] = o.Outlet 
        LEFT JOIN OutletMasterT AS o1 
             ON  c.[To PostCode] = o1.Postcode 
             AND c.[To Outlet] = o1.Outlet
        ORDER BY c.totalvalue DESC;")

SQL：

if object_id ('tempdb..#CustomerCodingChanges') is not null drop table #CustomerCodingChanges

SELECT c.CustID as ID
     , c.ReverseSupplier as Supplier
     , c.ReverseCustomerCode as Code
     , c.Name
     , c.Address
     , c.[From Postcode]
     , c.[From Outlet]
     , c.[From OutletName]
     , o.FullAddress AS [From Address]
     , c.[To Postcode]
     , c.[To Outlet]
     , c.[To OutletName]
     , o1.FullAddress AS [To Address]
     , CAST(c.TotalUnits AS VARCHAR(MAX)) AS [Total Units]
     , '$'+CAST(c.TotalValue AS VARCHAR(MAX)) AS [Total Value]
     , '' AS Checked 
     , c.CustRecActive as Active

INTO #CustomerCodingChanges
FROM #CustomerCorrectionSummary AS c
LEFT JOIN ndf_061.IRGMaster.dbo.OutletMaster AS o 
     ON  c.[From Postcode] = o.postcode 
     AND c.[From Outlet] = o.outlet
LEFT JOIN ndf_061.IRGMaster.dbo.OutletMaster AS o1 
     ON  c.[To Postcode] = o1.postcode 
     AND c.[To Outlet] = o1.outlet
ORDER BY c.totalvalue DESC;

CustomerCorrectionSummaryT 和 OutletMasterT 的两个数据框在 R 和 SQL Server 中具有相同数量的结果，所以我不知道为什么它不会在 R 和 SQL Server 中显示相同数量的结果。在 SQL Server 查询中返回 22 行，而在 R 中我只得到一个都正确的行。 R 的 sqldf 并没有显示全部。我认为这与我的 left join 函数有关，但我真的不知道。如果您需要更多信息，请告诉我！

Answer 1

如果您仔细查看，这两个查询并不相同。值得注意的是，您在第一个查询 MAX 中调用了一个聚合，而您在第二个查询中没有调用该聚合。因此，第一个查询是一个聚合查询，它通过返回聚合（SUM、COUNT、AVG、MIN、MAX 等）在指定列上的各种分组来折叠单元级行。

此外，由于您没有包含GROUP BY 子句，因此在标准 ANSI SQL 中，第一个查询应该会引发错误。不幸的是，对于 SQL 的新手来说，SQLite 和 MySQL 等方言允许这种语法。而 R 的 sqldf 默认运行 SQLite。事实上，如果您在 SQL Server 中运行第一个查询，它会出错，因为您通过在 GROUP BY 子句中未包含非聚合列（尽管它们在 SELECT 子句中指定）运行了不正确的聚合查询。

仔细观察，您似乎正试图在 SQL Server 中翻译以下表达式

CAST(c.TotalUnits AS VARCHAR(MAX)) AS [Total Units]

这与第一个查询中的尝试行不同：

MAX(CAST(c.TotalUnits AS VARCHAR)) AS [Total Units]

SQLite 实际上不维护maximum length in VARCHAR 类型。所以只需保持原始行没有聚合或大小限制：

CAST(c.TotalUnits AS VARCHAR) AS [Total Units]