SQL 服务器：对部分空数据使用聚合

Question

我的一些代码有问题，我需要一些帮助。下面的附件是从 PowerApps 进入 SQL 数据库的 table：


+-----------------------+-----------------+-------------+-----------+------------+------+-----------+
|          ID           | ProductionOrder | Workstation |  Status   |    Date    | Pass | TotalTime |
+-----------------------+-----------------+-------------+-----------+------------+------+-----------+
| 60127429_1000_1_S035  |          108100 | S035        | Completed | 05/13/2020 | No   |       121 |
| 60128523_1000_1_S120  |          884171 | S120        | Completed | 05/13/2020 | Yes  |        18 |
| 60128523_13000_1_S080 |          884172 | S080        | Completed | 05/13/2020 | No   |        26 |
| 60128523_13000_1_S090 |          884172 | S090        | Completed | 05/13/2020 | No   |        19 |
| 60128523_13000_1_S120 |          884172 | S120        | Completed | 05/13/2020 | No   |        16 |
| 60128523_28000_1_S070 |          884173 | S070        | Completed | 05/13/2020 | No   |        10 |
| 60128523_28000_1_S080 |          884173 | S080        | Completed | 05/13/2020 | No   |        16 |
| 60128523_28000_1_S090 |          884173 | S090        | Completed | 05/13/2020 | No   |        21 |
| 60128523_28000_1_S120 |          884173 | S120        | Completed | 05/13/2020 | No   |        17 |
| 60128523_47000_1_S080 |          884174 | S080        | Completed | 05/13/2020 | Yes  |        10 |
| 60128523_47000_1_S090 |          884174 | S090        | Completed | 05/13/2020 | Yes  |        50 |
| 60128523_47000_1_S120 |          884174 | S120        | Completed | 05/13/2020 | Yes  |        19 |
| 60128569_6000_1_S080  |          884698 | S080        | Completed | 05/13/2020 | Yes  |         9 |
| 60128569_6000_1_S090  |          884698 | S090        | Completed | 05/13/2020 | Yes  |        47 |
| 60128569_6000_1_S120  |          884698 | S120        | Completed | 05/13/2020 | Yes  |        54 |
| 60129071_1000_1_S070  |          105597 | S070        | Completed | 05/13/2020 | No   |       137 |
| 60129071_1000_1_S080  |          105597 | S080        | Completed | 05/13/2020 | No   |        12 |
| 60129071_1000_1_S090  |          105597 | S090        | Completed | 05/13/2020 | No   |        21 |
| 60129071_23000_1_S070 |          105598 | S070        | Completed | 05/13/2020 | Yes  |        50 |
| 60129071_23000_1_S080 |          105598 | S080        | Completed | 05/13/2020 | Yes  |        12 |
| 60129071_23000_1_S090 |          105598 | S090        | Completed | 05/13/2020 | Yes  |        29 |
| 60129071_23000_1_S120 |          105598 | S120        | Completed | 05/13/2020 | Yes  |        20 |
| 60129071_23025_1_S020 |          105600 | S020        | Completed | 05/13/2020 | Yes  |        21 |
| 60129151_2000_1_S090  |          104974 | S090        | Completed | 05/13/2020 | No   |        26 |
| 60129151_2000_1_S120  |          104974 | S120        | Completed | 05/13/2020 | No   |        24 |
+-----------------------+-----------------+-------------+-----------+------------+------+-----------+

接下来，我将创建一个视图，通过对每个工作站进行分组来获取一些指标。这是代码：

with cteNow AS 
(
    SELECT CASE WHEN 60 * DATEPART(HOUR, GETUTCDATE()) + DATEPART(MINUTE, GETUTCDATE()) > 60 * 21
            THEN CONVERT(float,(16 - 7) * 60) --minutes in (4 PM - 7 AM) * 60 minutes/hour
        ELSE
            CONVERT(float,60 * DATEPART(HOUR, GETUTCDATE()) + DATEPART(MINUTE, GETUTCDATE()) - ((7 + 5) * 60))
        END as MinutesToday
        , CONVERT(varchar(50), DATEADD(HOUR, -5, GETUTCDATE()), 101) as TargetDate 
)
SELECT dbs.Workstation, 

100.0 - ISNULL(SUM(CASE WHEN dbs.Pass = 'Yes' AND Status = 'Completed' THEN 1 ELSE 0 END) / NULLIF(SUM(CASE WHEN dbs.Status = 'Completed' THEN 1 ELSE 0 END), 0), 0) AS RFT

FROM [Booms DBS] AS dbs LEFT JOIN [Booms Delay Entry] AS del ON dbs.ID = del.ID INNER JOIN cteNow AS N ON N.TargetDate = dbs.[Date]
GROUP BY dbs.Workstation, MinutesToday

这样做的目的是为每个站点测量 "Right First Time"。每当 Pass 列中出现 "Yes" 值时，它就是对 RFT 的命中。例如，如果给定站点中总共有十个构建，其中五个在 Pass 列中具有 "Yes"，则为 50% RFT。

但是，这总是给我 RFT 列的结果“100”。我添加了所有这些 NULLIF 和 ISNULL 警告，因为有时 Pass 列可能为空，并且在某些日子里，可能根本没有任何数据，尤其是在早上。我假设这就是问题所在。我没有正确计算吗？如果我不包含那些空值检查函数，我想不出任何其他方法来避免出错或出现被零除错误。

提前致谢。

Answer 1

The point of this is to measure "Right First Time" for each station. Every time there is a "Yes" value in the Pass column, it is a hit to RFT. For example, if there is ten total builds in a given station and five of them have "Yes" in the Pass column, that is 50% RFT.

您似乎需要条件聚合。我不明白您现有查询中日期算法的目的，但这会按照您的描述进行：

select 
    workstation, 
    avg(case when statys = 'Completed' and pass = 'Yes' then 1.0 else 0 end) rft_rate
from dbs.workstation
group by workstation

这为每个 workstation 提供一条记录，数值介于 0 和 1 之间，表示在 pass 列中具有值 'Yes' 的记录的比率。[=15] =]

SQL 服务器：对部分空数据使用聚合

SQL Server : using aggregations on partially null data

sql

tsql

sql-server

group-by

powerapps