如何在不添加到 GROUP BY 的情况下计算不同的值

Question

我有一个数据集，其中包含每个“作业”记录的名称，以及作业是通过还是失败。我想在一行中显示名称、作业数量、通过的数量和失败的数量。

我正在对名称进行分组，并在名称上使用 COUNT 来计算作业总数，这工作正常，但如果不将它们添加到导致数据的 GROUP BY 子句中，我无法显示通过的数量和失败的数量再次分开。

SELECT  I.Name, Count(I.Name) As NumberOfJobs,
CASE WHEN WI.resultTypeID = 1 THEN COUNT(WI.resultTypeID) END AS [Passed],
CASE WHEN WI.resultTypeID = 2 THEN COUNT(WI.resultTypeID) END AS [Failed],
FROM DB.DBO.People AS I 
INNER JOIN DB2.dbo.Jobs AS WI  ON I.JOBID = WI.JOBID
GROUP BY I.Name, wi.resultTypeID

    +-----------+-----------+--------+--------+
    |   Name    | NumofJobs | Passed | Failed |
    +-----------+-----------+--------+--------+
    | Dale Test |         2 | 2      | NULL   |
    | Dale Test |         2 | NULL   | 2      |
    +-----------+-----------+--------+--------+

当我将 ResultTypeID 添加到 GROUP BY 时会发生这种情况，但我想要这个：

    +-----------+-----------+--------+--------+
    |   Name    | NumofJobs | Passed | Failed |
    +-----------+-----------+--------+--------+
    | Dale Test |         4 |      2 |      2 |
    +-----------+-----------+--------+--------+

有没有办法做到这一点？

Answer 1

您需要条件聚合。 case 表达式是聚合函数的参数：

SELECT I.Name, Count(*) As NumberOfJobs,
       SUM(CASE WHEN WI.resultTypeID = 1 THEN 1 ELSE 0 END) AS [Passed],
       SUM(CASE WHEN WI.resultTypeID = 2 THEN 1 ELSE 0 END) AS [Failed],
FROM DB.DBO.People I INNER JOIN
     DB2.dbo.Jobs WI
     ON I.JOBID = WI.JOBID
GROUP BY I.Name;

我猜wi.resultTypeID 不是NULL，所以我将COUNT() 替换为SUM()，因为在这种情况下我更喜欢SUM()。

Answer 2

您无需将查询按wi.resultTypeID 分组。

只需从 group by 语句中删除 wi.resultTypeID 并将其放入聚合函数中：

SELECT  I.Name, Count(I.Name) As NumberOfJobs,
SUM(CASE WHEN WI.resultTypeID = 1 THEN 1 ELSE 0 END) AS [Passed],
SUM(CASE WHEN WI.resultTypeID = 2 THEN 1 ELSE 0 END) AS [Failed],
FROM DB.DBO.People AS I 
INNER JOIN DB2.dbo.Jobs AS WI  ON I.JOBID = WI.JOBID
GROUP BY I.Name