Sql 将“联合所有”与“分组”相结合

在这个问题上，我真的需要一些帮助

我有一张产品表

和一个事务表

和一个CCTransactions表

我需要一个查询，该查询可以联合所有在给定日期范围内的事务和CCTransactions，然后将它们分组为Product.Id、Product.Title、SUMQuantity和SUMTotalAmount

以下是我到目前为止的情况。它不会编译，但我会包括在内，因为人们通常希望看到你尝试过什么。它告诉我：

Msg 8120，16级，状态1，第5行 列“Products.Id”在选择列表中无效，因为它未包含在聚合函数或GROUP BY子句中。 Msg 209，16级，状态1，第14行 列名称“Id”不明确

我理解这些错误，但不确定修复它们的最佳方法。在我加入工会之前，这个问题似乎起了作用

DECLARE @dtStart DATE = '2016-08-01';
DECLARE @dtEnd DATE = '2016-08-31';

SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM Transactions t
INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
UNION ALL
SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM CCTransactions t
INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
GROUP BY p.Id, p.Title
ORDER BY Title

您需要两个单独的分组：

或子查询：

SELECT      X.id, X.Title, SUM(X.Quantity), SUM(X.TotalAmount)
FROM        (
                SELECT p.Id, p.Title, t.Quantity, t.TotalAmount
                FROM Transactions t
                INNER JOIN Products p ON t.ProductId = p.Id
                WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd

                UNION ALL

                SELECT p.Id, p.Title, t.Quantity, t.TotalAmount
                FROM CCTransactions t
                INNER JOIN Products p ON t.ProductId = p.Id
                WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
            ) X
GROUP BY    X.Id, X.Title

或者您可以重写整个查询：

;WITH
    AllTransaction AS
    (
        SELECT      ProductID, Timestamp, Quantity, TotalAmount
        FROM        Transaction
        UNION ALL
        SELECT      ProductID, Timestamp, Quantity, TotalAmount
        FROM        CCTransaction
    )

SELECT      p.Id, p.Ttitle, SUM(t.Quantity), SUM(t.TotalAmount)
FROM        AllTransaction      t
INNER JOIN  Products            p   ON t.ProductId = p.Id
WHERE       @dtStart <= t.[Timestamp] AND t.[TimeStamp] < DATEADD(DAY, 1, @dtEnd)
GROUP BY    p.Id, p.Title

---

关于cast。[TimeStamp]AS DATE的说明首先对2个表使用UNION ALL，并将结果连接到产品：

DECLARE @dtStart DATE = '2016-08-01';
DECLARE @dtEnd DATE = '2016-08-31';

SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM (
  SELECT ProductId, [TimeStamp], Quantity, TotalAmount FROM Transactions
  UNION ALL
  SELECT ProductId, [TimeStamp], Quantity, TotalAmount FROM CCTransactions
) t INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
GROUP BY p.Id, p.Title
ORDER BY Title

---

我喜欢这种方法。我正在确认我是否得到了我期望的数字，我得到了。是的，我喜欢这样的方式：在进行所有额外的联合和分组之前，先获取底层数据。不确定它是否执行得最好，但似乎最容易理解。您也可以尝试应用以下条件：WHERE[TimeStamp]>=@dtStart并将[TimeStamp]强制转换为DATE Yes，这将生成更长的查询。很想知道它对性能的影响。在两个表中的[TimeStamp]上都有索引，这可能会比外部WHERE条件更好。第一个1在联合之前有GROUP BY，与其他2不同。对于AVG，这会产生不同。对于SUM来说，它是否对SUM重要并不重要。对于事务和cctransactions中出现的任何内容，您都将获得具有相同IT、标题组合的多行，因为在进行联合之前，在没有任何聚合的情况下进行分组。您必须进行另一次求和/分组才能再次合并2。

SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM Transactions t
INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
GROUP BY p.Id, p.Title

UNION ALL

SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM CCTransactions t
INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
GROUP BY p.Id, p.Title

SELECT      X.id, X.Title, SUM(X.Quantity), SUM(X.TotalAmount)
FROM        (
                SELECT p.Id, p.Title, t.Quantity, t.TotalAmount
                FROM Transactions t
                INNER JOIN Products p ON t.ProductId = p.Id
                WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd

                UNION ALL

                SELECT p.Id, p.Title, t.Quantity, t.TotalAmount
                FROM CCTransactions t
                INNER JOIN Products p ON t.ProductId = p.Id
                WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
            ) X
GROUP BY    X.Id, X.Title

;WITH
    AllTransaction AS
    (
        SELECT      ProductID, Timestamp, Quantity, TotalAmount
        FROM        Transaction
        UNION ALL
        SELECT      ProductID, Timestamp, Quantity, TotalAmount
        FROM        CCTransaction
    )

SELECT      p.Id, p.Ttitle, SUM(t.Quantity), SUM(t.TotalAmount)
FROM        AllTransaction      t
INNER JOIN  Products            p   ON t.ProductId = p.Id
WHERE       @dtStart <= t.[Timestamp] AND t.[TimeStamp] < DATEADD(DAY, 1, @dtEnd)
GROUP BY    p.Id, p.Title

DECLARE @dtStart DATE = '2016-08-01';
DECLARE @dtEnd DATE = '2016-08-31';

SELECT p.Id, p.Title, SUM(t.Quantity), SUM(t.TotalAmount) AS Amount
FROM (
  SELECT ProductId, [TimeStamp], Quantity, TotalAmount FROM Transactions
  UNION ALL
  SELECT ProductId, [TimeStamp], Quantity, TotalAmount FROM CCTransactions
) t INNER JOIN Products p ON t.ProductId = p.Id
WHERE t.[TimeStamp] >= @dtStart AND CAST(t.[TimeStamp] AS DATE) <= @dtEnd
GROUP BY p.Id, p.Title
ORDER BY Title