Question

首先，我声称没有SQL的极端专业知识。

我要做的是根据几个不同的标准从SQL Server数据库中提取一些事件计数，并将它们放入SSRS中，在一个折线图中显示每个月的计数。我最初创建的两个查询都会下拉正确的数据，但SSRS不允许在一个图表上使用多个数据集。这是我被挂断的地方。我确信我可以进行某种子查询，将两个数据集合成一个只需将计数分成列的数据集，但到目前为止我所尝试的数据集合了整个计数。

我需要的数据是报告的日期，它是日期/时间值和发生的事件计数（无需汇总计数，或者不应该，因为SSRS将允许我按月对它们进行分组，这是报告应该是什么样的）匹配2个标准中的每一个。

以下是我正在使用的两个查询：

SELECT DISTINCT v_Incident.ReportedDate, COUNT(*) AS Count
FROM            v_Incident INNER JOIN
                         v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
WHERE        (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') AND (v_Incident.CustomerFirstName = N'BMC') AND 
                         (v_Incident.CustomerLastName = N'Impact Manager') AND (v_IncidentMaxWorkLog_Category.WORK_LOG_TYPE IN ('General Information', 'Incident Task / Action', 
                         'Working Log')) AND (v_Incident.ReportedDate >= '7/1/2014') AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
GROUP BY v_Incident.ReportedDate
FROM v_Incident

第二次查询

SELECT DISTINCT v_Incident.ReportedDate, COUNT(*) AS Count
FROM            v_Incident INNER JOIN
                         v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
WHERE        (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') AND (v_Incident.CustomerFirstName = N'BMC') AND 
                         (v_Incident.CustomerLastName = N'Impact Manager') AND (v_Incident.ReportedDate >= '7/1/2014') AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
GROUP BY v_Incident.ReportedDate
FROM v_Incident

正如你所看到的，我正在拉下报告的日期和计数。如果我只是将这两个语句放在一个语句中[select（query1）as filtered，（query2）as base]并将组移到外面，我得到结果，但它是整个每一行的每个值，以及大约10k行太多。

是否有一个简单的解决方案可以获取每个日期，并使用这两个或类似的查询进行基数和过滤的单一计数？非常感谢你提前！

Answer 1

根据我的想法，我将使用UNION并将联合查询视为subquery。

SELECT * 
FROM
    ((SELECT DISTINCT 
          v_Incident.ReportedDate, COUNT(*) AS Count, 'BASE' as [Designator]
      FROM            
          v_Incident 
      INNER JOIN
          v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
      WHERE        
         (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') 
         AND (v_Incident.CustomerFirstName = N'BMC') 
         AND (v_Incident.CustomerLastName = N'Impact Manager') 
         AND (v_IncidentMaxWorkLog_Category.WORK_LOG_TYPE IN ('General Information', 'Incident Task / Action', 'Working Log')) 
         AND (v_Incident.ReportedDate >= '7/1/2014') 
         AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
     GROUP BY 
         v_Incident.ReportedDate
FROM v_Incident)

UNION ALL

(SELECT DISTINCT v_Incident.ReportedDate, COUNT(*) AS Count, 'Filtered' as [Designator]
FROM            v_Incident INNER JOIN
                         v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
WHERE        (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') AND (v_Incident.CustomerFirstName = N'BMC') AND 
                         (v_Incident.CustomerLastName = N'Impact Manager') AND (v_Incident.ReportedDate >= '7/1/2014') AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
GROUP BY v_Incident.ReportedDate
FROM v_Incident)) ORDER BY ReportedDate, Designator

Answer 2

select isnull(a.ReportedDate, b.ReportedDate) ReportedDate, 
       a.Count ACount, b.Count BCount
from
(
    SELECT DISTINCT v_Incident.ReportedDate, COUNT(*) AS Count
    FROM            v_Incident INNER JOIN
                             v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
    WHERE        (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') AND (v_Incident.CustomerFirstName = N'BMC') AND 
                             (v_Incident.CustomerLastName = N'Impact Manager') AND (v_IncidentMaxWorkLog_Category.WORK_LOG_TYPE IN ('General Information', 'Incident Task / Action', 
                             'Working Log')) AND (v_Incident.ReportedDate >= '7/1/2014') AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
    GROUP BY v_Incident.ReportedDate
    FROM v_Incident
) a
full outer join
(
    SELECT DISTINCT v_Incident.ReportedDate, COUNT(*) AS Count
    FROM            v_Incident INNER JOIN
                             v_IncidentMaxWorkLog_Category ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
    WHERE        (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%') AND (v_Incident.CustomerFirstName = N'BMC') AND 
                             (v_Incident.CustomerLastName = N'Impact Manager') AND (v_Incident.ReportedDate >= '7/1/2014') AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
    GROUP BY v_Incident.ReportedDate
    FROM v_Incident
) b on b.ReportedDate = a.ReportedDate

这会处理您的每个原始查询，并将其作为表格进行处理。到ReportedDate连接它们的第三个查询。它使用Full Outer Join，因此如果缺少b，仍将检索a，如果缺少a，则仍然会报告b。这种方式有3种组合：

a在那里，b为空
a和b在那里
a为空，b就在那里因此，要在结果中获取ReportedDate，请使用isnull来显示a.ReportedDate，并且该值为null（上面的情况3）使用b.ReportedDate。

请记住，在任何SQL Select中，您都可以使用Select查询替换表名。

Answer 3

您的查询非常相似。看起来您有一个总计数，然后是一些具有特定工作日志类型的记录的部分计数。这是一个返回date，BaseCount和'WorkLogFilteredCount'的查询.SSRS应该能够将日期作为轴，而其他两列作为单独的系列非常容易用于您的图表。

SELECT  v_Incident.ReportedDate
        , COUNT(*) AS BaseCount
        , SUM(CASE WHEN v_IncidentMaxWorkLog_Category.WORK_LOG_TYPE IN ('General Information', 'Incident Task / Action', 'Working Log') THEN 1 ELSE 0 END) AS WorkLogFilteredCount

FROM    v_Incident
        INNER JOIN v_IncidentMaxWorkLog_Category
            ON v_Incident.IncidentID = v_IncidentMaxWorkLog_Category.Incident_Number
WHERE   (v_Incident.Summary LIKE N'Intelligent Incident for Impacted CI%')
        AND (v_Incident.CustomerFirstName = N'BMC')
        AND (v_Incident.CustomerLastName = N'Impact Manager')
        AND (v_Incident.ReportedDate >= '7/1/2014')
        AND (v_Incident.ReportedDate >= DATEADD(year, - 1, GETDATE()))
GROUP BY v_Incident.ReportedDate
FROM v_Incident

TSQL 2查询1

3 个答案: