不能缺少显示零的天数

Posted 2023-02-24

【中文标题】不能缺少显示零的天数

【英文标题】：Can't get missing days to show zero

【发布时间】：2019-12-25 23:16:13

【问题描述】：

这是我写的 - 我试图让没有交易的日期只显示该字段的零。我已经去谷歌并和我认识的每个人交谈，这应该可行 - 但不是：？

IF OBJECT_ID('tempdb..#D') IS NOT NULL
    DROP TABLE #D;

CREATE TABLE #D
(
    [Date] [DATE] NOT NULL,
    PRIMARY KEY CLUSTERED ([Date] ASC)
);

DECLARE @CurrentDate DATE = GETDATE() - 95;
DECLARE @EndDate DATE = GETDATE();

WHILE @CurrentDate < @EndDate
BEGIN
    INSERT INTO #D
    (
        [Date]
    )
    SELECT DATE = @CurrentDate;
    SET @CurrentDate = DATEADD(DD, 1, @CurrentDate);
END;

--SELECT * FROM #D ORDER BY  Date DESC


SELECT ITEM2_NUM AS 'Part Number',
       COALESCE(SUM(TRANS_QTY), 0) 'Daily Usage',
       D.Date AS 'Usage Date'
FROM #D AS D
    LEFT OUTER JOIN dbo.W_INVENTORY_TRANS_F IT
        ON IT.GL_DT = D.Date
    LEFT JOIN dbo.W_BU_ITEM_D BI
        ON BI.BUSINESS_UNIT_WID = IT.BUSINESS_UNIT_WID
           AND BI.ITEM_WID = IT.ITEM_WID
WHERE DOCUMENT_TYPE_WID = 22
      AND D.Date >= DATEADD(yy, -1, GETDATE())
      AND IT.BUSINESS_UNIT_WID = '837'
      AND IT.ITEM2_NUM = '10111'
      AND BI.STOCKING_TYPE = 'P'
GROUP BY ITEM2_NUM,
         D.Date
ORDER BY D.Date DESC;

以下是我得到的结果：

Part Number Daily Usage Usage Date
10111   -331.0000   2019-08-19
10111   -2617.0000  2019-08-16
10111   -418.0000   2019-08-15
10111   -471.0000   2019-08-14
10111   -1158.0000  2019-08-13
10111   -766.0000   2019-08-12
10111   -1385.0000  2019-08-09

这就是我想要的：

Part Number Daily Usage Usage Date
10111   -331    8/19/2019
10111     0     8/18/2019
10111     0     8/17/2019
10111   -2617   8/16/2019
10111   -418    8/15/2019
10111   -471    8/14/2019
10111   -1158   8/13/2019
10111   -766    8/12/2019
10111     0     8/11/2019
10111     0     8/10/2019
10111   -1385   8/9/2019

【问题讨论】：

请在代码问题中给出minimal reproducible example--剪切&粘贴&可运行代码；具有期望和实际输出的示例输入（包括逐字错误消息）；标签和版本；明确的规范和解释。这包括您可以提供的最少代码，即您显示的代码可以通过您显示的代码扩展为不正常。 （调试基础。） PS如果你不努力清楚地说出你想要什么，那么你就无法理解、推理、交流或搜索。使用足够多的单词、句子和对部分示例的引用来清楚完整地表达你的意思。

了解 LEFT JOIN ON 返回的内容：INNER JOIN ON 行 UNION ALL 不匹配的左表行，由 NULL 扩展。作为 OUTER JOIN 的一部分，始终知道您想要什么 INNER JOIN。 WHERE 或 INNER JOIN ON 在 OUTER JOIN ON 删除任何由 NULL 扩展的行后，需要右 [sic] 表列不为 NULL，即只留下 INNER JOIN ON 行，即“将 OUTER JOIN 转换为 INNER JOIN”。你有那个。

这是一个常见问题解答。在考虑发布之前，请始终在谷歌上搜索任何错误消息和许多清晰、简洁和准确的问题/问题/目标的措辞，带或不带您的特定字符串、名称和行号以及带和不带“site:***.com”并阅读很多点击和答案。如果您发布问题，请使用一个短语作为标题。请参阅How to Ask 和投票箭头鼠标悬停文本。

Left Outer Join doesn't return all rows from my left table?的可能重复

【参考方案1】：

您的where 子句实际上将您的left join 更改为inner join。 这是因为left join 为在正确表中找不到的行返回空值，而null 与任何东西相比总是返回unknown，SQL Server 将其解释为false。

您需要做的是将所有引用右侧表的条件移到 on 子句中。 此外，您正在按 ITEM2_NUM 分组，这没有意义，因为它已经在 where 子句中被过滤掉了。

这应该是您的代码的工作版本（DOCUMENT_TYPE_WID 列除外，我不知道它属于哪个表）：

SELECT '10111' AS 'Part Number',
       COALESCE(SUM(TRANS_QTY), 0) 'Daily Usage',
       D.Date AS 'Usage Date'
FROM #D AS D
    LEFT OUTER JOIN dbo.W_INVENTORY_TRANS_F IT
        ON IT.GL_DT = D.Date
        AND IT.BUSINESS_UNIT_WID = '837'
        AND IT.ITEM2_NUM = '10111'
        -- I had to guess where this column belongs, it might be on the BI table
        AND IT.DOCUMENT_TYPE_WID = 22
    LEFT JOIN dbo.W_BU_ITEM_D BI
        ON BI.BUSINESS_UNIT_WID = IT.BUSINESS_UNIT_WID
        AND BI.ITEM_WID = IT.ITEM_WID
        AND BI.STOCKING_TYPE = 'P'
WHERE D.Date >= DATEADD(yy, -1, GETDATE())
GROUP BY D.Date
ORDER BY D.Date DESC;

【参考方案2】：

Zohar 是 100% 正确的 - 如果我最终使用的最终代码会在下面的某个时间点结束。

IF OBJECT_ID('tempdb..#D') IS NOT NULL
    DROP TABLE #D;

CREATE TABLE #D
(
    [Date] [DATE] NOT NULL,
    PRIMARY KEY CLUSTERED ([Date] ASC)
);

DECLARE @CurrentDate DATE = GETDATE() - 95;
DECLARE @EndDate DATE = GETDATE();

WHILE @CurrentDate < @EndDate
BEGIN
    INSERT INTO #D
    (
        [Date]
    )
    SELECT DATE = @CurrentDate;
    SET @CurrentDate = DATEADD(DD, 1, @CurrentDate);
END;

IF OBJECT_ID('tempdb..#DailyUsage') IS NOT NULL
    DROP TABLE #DailyUsage;

CREATE TABLE #DailyUsage
(
    BU_ITEM_WID INT,
    TRANS_QTY DECIMAL(15, 4),
    GL_DT DATETIME
);
INSERT INTO #DailyUsage
(
    BU_ITEM_WID,
    TRANS_QTY,
    GL_DT
)
SELECT BU_ITEM_WID,
       TRANS_QTY AS 'Daily Usage',
       GL_DT
FROM dbo.W_INVENTORY_TRANS_F
WHERE 1 = 1
      AND DOCUMENT_TYPE_WID = 22
      AND GL_DT >= GETDATE() - 95;

CREATE NONCLUSTERED INDEX [IX_DailyUsage]
ON #DailyUsage (BU_ITEM_WID)
INCLUDE
(
    TRANS_QTY,
    GL_DT
);

WITH cte_ItemNums
AS (SELECT DISTINCT
        IT.ITEM2_NUM,
        IT.BU_ITEM_WID
    FROM W_INVENTORY_TRANS_F IT
        LEFT JOIN dbo.W_BU_ITEM_D BI
            ON BI.BUSINESS_UNIT_WID = IT.BUSINESS_UNIT_WID
               AND BI.ITEM_WID = IT.ITEM_WID
               AND BI.STOCKING_TYPE = 'P'
    WHERE 1 = 1
          AND IT.DOCUMENT_TYPE_WID = 22
          AND IT.BUSINESS_UNIT_WID = '837'
   )
SELECT IT1.ITEM2_NUM,
       COALESCE(SUM(IT2.TRANS_QTY), 0) AS 'Daily Usage',
       D.Date AS 'Usage Date'
FROM #D AS D
    LEFT OUTER JOIN cte_ItemNums IT1
        ON 1 = 1
    LEFT OUTER JOIN #DailyUsage IT2
        ON IT2.GL_DT = D.Date
           AND IT1.BU_ITEM_WID = IT2.BU_ITEM_WID
GROUP BY IT1.ITEM2_NUM,
         D.Date;