在 SQL Server 2014 中计算每个业务部门的员工呼叫比例

Question

我正在制定业务规则，以确定每 30 分钟间隔活跃的呼叫中心代理的数量。

我们有两种类型的代理：计费和技术运营。

我们最近让计费代理接听技术电话，反之亦然。

我编写了以下代码，它返回每个代理、每个间隔和每个呼叫类型的呼叫计数：

SELECT 
    HAGENT.[ROW_DATE],
    CASE 
       WHEN LEN([STARTTIME]) = 1 THEN '00:00'
       WHEN LEN([STARTTIME]) = 2 THEN '00:30'
       WHEN LEN([STARTTIME]) = 3 THEN (LEFT([STARTTIME],1) + ':' + RIGHT([STARTTIME],2))
       WHEN LEN([STARTTIME]) = 4 THEN (LEFT([STARTTIME],2) + ':' + RIGHT([STARTTIME],2)) 
    END AS THETIME,
    ([LOGID]) AS AGENTS,
    COUNT(HAGENT.ACDCALLS) AS CALL_COUNT,
    CASE 
       WHEN SPLITMATRIX.[Business_Unit] LIKE '%Billing%' THEN 'CDN Billing'
       WHEN SPLITMATRIX.[Business_Unit] LIKE '%TechOps%' THEN 'CDN Tech Ops' 
    END AS Bunit
FROM 
    [KPI_DATA].[dbo].[DM_CMS_HAGENT] HAGENT
LEFT JOIN 
    [GEMDB].[dbo].[T_PRD_REF_xAGENT_AGENTLIST] AGENTLIST 
        ON AGENTLIST.asp_raw_cmsID1 = HAGENT.LOGID  COLLATE DATABASE_DEFAULT 
        AND HAGENT.ROW_DATE BETWEEN AGENTLIST.[asp_raw_START_SDEPT] AND AGENTLIST.asp_raw_STOP_SDEPT
        AND HAGENT.ROW_DATE BETWEEN AGENTLIST.[asp_raw_START_TEAM] AND AGENTLIST.[asp_raw_STOP_TEAM]
LEFT JOIN 
    [MATRIX_DATA].[dbo].[REF_CMS_SPLIT_MATRIX] SPLITMATRIX
        ON SPLITMATRIX.SPLIT = HAGENT.SPLIT
WHERE
    ROW_DATE >= '2017-07-01' 
    AND [ACDCALLS] != 0 
    AND (SPLITMATRIX.[Business_Unit] LIKE '%Billing%' 
         OR SPLITMATRIX.[Business_Unit] LIKE '%TechOps%')
GROUP BY 
    GROUPING SETS (
         (ROW_DATE, STARTTIME,([LOGID]), SPLITMATRIX.[Business_Unit]),
         (ROW_DATE, STARTTIME,([LOGID]))
    )
ORDER BY 
    ROW_DATE DESC, THETIME DESC, [LOGID] DESC

输出如下所示：

SQL Output

由于 AGENTS 是代理的唯一 logid，Bunit 是所接电话的类型（当 Bunit 为 NULL 时，它是该时间间隔内代理所接电话的总数）。

从这个输出中，我想计算每个间隔和呼叫类型的活跃员工数。

例如，如果代理接听了 2 个计费电话和 1 个技术运营电话，我会认为该代理的 2/3 是计费代理，而该代理的 1/3 是技术运营代理。

最后，我想对每个呼叫类型、每个间隔的座席比率的总和进行分组。

我尝试使用 PIVOT 函数，但没有成功。

我的预期结果是这样的：

https://i.stack.imgur.com/oKiUl.jpg

有什么帮助吗？

Answer 1

您可以在这种情况下使用 PIVOT 和 UNPIVOT

样本数据

CREATE TABLE t_bu
(row_date DATE,
 thetime TIME,
 agents INT,
 call_count INT,
 bunit VARCHAR(20)
)

INSERT INTO t_bu VALUES
('8/17/2017','9:30',1210893,1,'CDN Billing'),
('8/17/2017','9:30',1210893,1,NULL),
('8/17/2017','9:30',1210870,2,'CDN Tech Ops'),
('8/17/2017','9:30',1210870,2,NULL),
('8/17/2017','9:30',1210858,3,'CDN Billing'),
('8/17/2017','9:30',1210858,3,NULL),
('8/17/2017','9:30',1210857,1,'CDN Tech Ops'),
('8/17/2017','9:30',1210857,1,NULL),
('8/17/2017','9:30',1210849,1,'CDN Billing'),
('8/17/2017','9:30',1210849,1,NULL),
('8/17/2017','9:30',1210832,2,'CDN Billing'),
('8/17/2017','9:30',1210832,1,'CDN Tech Ops'),
('8/17/2017','9:30',1210832,3,NULL),
('10/21/2017','10:30',1210857,3,'CDN Tech Ops'),
('10/21/2017','10:30',1210857,3,NULL),
('10/21/2017','12:30',1210857,3,'CDN Tech Ops'),
('10/21/2017','12:30',1210857,3,NULL),
('10/21/2017','12:30',1210849,4,'CDN Billing'),
('10/21/2017','12:30',1210849,4,NULL),
('10/21/2017','12:30',1210832,3,'CDN Billing'),
('10/21/2017','12:30',1210832,5,'CDN Tech Ops'),
('10/21/2017','12:30',1210832,8,NULL);

SQL 查询（动态）

DECLARE @cols  AS NVARCHAR(MAX),
        @query AS NVARCHAR(MAX)

SET @cols = STUFF((SELECT DISTINCT ',' + QUOTENAME(CASE WHEN t.bunit IS NULL THEN 'Total' ELSE t.bunit END) 
                     FROM t_bu t -- use your table name
                      FOR XML PATH(''), TYPE
                  ).value('.', 'NVARCHAR(MAX)'),1,1,'')

SET @query = 'WITH t_comp AS
              (SELECT row_date, thetime, SUM(CAST([CDN Billing] AS FLOAT)/CAST([Total] AS FLOAT)) [CDN Billing], SUM(CAST([CDN Tech Ops] AS FLOAT)/CAST([Total] AS FLOAT)) [CDN Tech Ops]
                 FROM (SELECT agents, row_date, thetime, ' + @cols + 
             '           FROM (SELECT agents, 
                                      row_date,
                                      thetime,
                                      CASE WHEN bunit IS NULL THEN ''Total'' ELSE bunit END AS bunit,
                                      call_count
                                 FROM t_bu -- use your table name
                              ) x
                        PIVOT (MAX(call_count)
                          FOR bunit in (' + @cols + ')) p 
                      ) a
                GROUP BY row_date, thetime
               )
               SELECT row_date, thetime, call_type, active_agents
                 FROM t_comp
               UNPIVOT (active_agents
                   FOR call_type IN ([CDN Billing], [CDN Tech Ops])) u'

EXECUTE(@query)
 GO

结果

row_date    thetime             call_type       active_agents
2017-08-17  09:30:00.0000000    CDN Billing     3.66666666666667
2017-08-17  09:30:00.0000000    CDN Tech Ops    2.33333333333333
2017-10-21  10:30:00.0000000    CDN Tech Ops    1
2017-10-21  12:30:00.0000000    CDN Billing     1.375
2017-10-21  12:30:00.0000000    CDN Tech Ops    1.625