Question

我可能在这里遗漏了一些简单的东西。我有第一张表：

CREATE TABLE [Orgnzs] (
    [id] INT NOT NULL IDENTITY(1,1) PRIMARY KEY,
    [nm] NVARCHAR(256)
);

然后还有几个表都是这样创建的（都具有相同的结构）：

-- WLog_* tables are all created as such
CREATE TABLE [WLog_1] (
    [id] BIGINT NOT NULL IDENTITY(1,1) PRIMARY KEY,
    [huid] UNIQUEIDENTIFIER,
    [dtin] BIGINT,
    [dtout] BIGINT,
    [cnm] NVARCHAR(15),
    [batt] TINYINT,
    [pwrop] TINYINT,
    [pst] INT,
    [flgs] INT,
    [ppocs] NVARCHAR(1024),
    [ppocu] NVARCHAR(1024),
    [por] NVARCHAR(1024)
);

和一组类似的表，没有上表中的最后3列：

-- ULog_* tables are all created as such
CREATE TABLE [ULog_1] (
    [id] BIGINT NOT NULL IDENTITY(1,1) PRIMARY KEY,
    [huid] UNIQUEIDENTIFIER,
    [dtin] BIGINT,
    [dtout] BIGINT,
    [cnm] NVARCHAR(15),
    [batt] TINYINT,
    [pwrop] TINYINT,
    [pst] INT,
    [flgs] INT
);

我的目标是从任意一组WLog_ *和ULog_ *表中选择记录，并通过可管理的元素数量（用于页面布局）限制它，我还需要知道找到的记录总数。

所以我这样选择：

SELECT  b.[id] AS evtID,
        b.[huid] as huid,
        b.[dtin] as dtin,
        b.[dtout] as dtout,
        b.[cnm] as cnm,
        b.[batt] as batt,
        b.[pwrop] as pwrop,
        b.[pst] as pst,
        b.[flgs] as flgs,
        b.[ppocs] as ppocs,
        b.[ppocu] as ppocu,
        b.[por] as por,
        b.[orgID] as orgID,
        b.[wLg] as wLg,
        orgz.[nm] as orgNm
        , COUNT_BIG(*) as allRecordsFound
 FROM (
    -- next also specify the column(s) to sort by
    SELECT *, ROW_NUMBER() OVER (ORDER BY [dtin], [cnm] ASC) AS rw FROM (
        SELECT *, 1 AS orgID, 1 AS wLg
          FROM [WLog_1]
        UNION ALL
        SELECT *, 2 AS orgID, 1 AS wLg
          FROM [WLog_2]
        UNION ALL
        SELECT *, NULL AS [ppocs], NULL AS [ppocu], NULL AS [por], 1 AS orgID, 0 AS wLg
          FROM [ULog_1]
    ) a
    WHERE [pst]&1=1 OR [pst]=67
) b 
LEFT JOIN [Orgnzs] AS orgz ON orgID=orgz.[id]
WHERE rw >= 2 AND rw <= 4       -- restrict for a page only

，遗憾的是COUNT_BIG(*) as allRecordsFound行失败，出现以下错误：

专栏＆＃39; b.id＆＃39;在选择列表中无效，因为它不是包含在聚合函数或GROUP BY子句中。

我还没有使用SQL Server一段时间，有人可以建议我在这里缺少什么吗？

PS。出于测试目的，我made a Fiddle尝试一下。

Answer 1

请改用：

count(*) over() as allRecordsFound

您可以在select语句中混合窗口聚合函数，而不用分组。

COUNT没有GROUP BY子句问题

1 个答案: