Question

这里有另一个新手问题，

我有以下选择：

SELECT    
    DATEADD(dd, 0, DATEDIFF(dd, 0, CIC.StatusDateTime)) AS ST_Date,
    CIC.UserId,
    MIN(CIC.StatusDateTime) as Beginning,
    MAX(CIC.EndDateTime) as ending,
    DATEPART(hour,MAX(CIC.EndDateTime) - MIN(CIC.StatusDateTime)) * 3600 
     + DATEPART(minute, MAX(CIC.EndDateTime) - MIN(CIC.StatusDateTime)) * 60 
     + DATEPART(second,MAX(CIC.EndDateTime) - MIN(CIC.StatusDateTime)) AS Login_Time, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 5 THEN StateDuration ELSE 0 END) AS Lunch, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 3 THEN StateDuration ELSE 0 END) AS Break_, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 4 THEN StateDuration ELSE 0 END) AS Coaching, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 10 THEN StateDuration ELSE 0 END) AS SYS_Down, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 7 THEN StateDuration ELSE 0 END) AS Training, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 6 THEN StateDuration ELSE 0 END) AS Meeting, 
    SUM(CASE WHEN RS.AGT_STATE_GRP = 11 THEN StateDuration ELSE 0 END) AS Default_
FROM dbo.CIC_AgentActivityLog_TMP AS CIC
INNER JOIN dbo.REF_AGT_STATES AS RS 
ON CIC.StatusKey = RS.AGT_STATE_DESC
Where CIC.StatusDateTime >='2011-09-01'
GROUP BY CIC.StatusDateTime, CIC.UserId, CIC.EndDateTime
ORDER BY ST_Date,UserID,StatusDateTime;

使用它，我的结果似乎根本不是分组。我想要的是一行包含date，userid，firstlogin，lastlogin，totallogintime，然后是一堆代码，其持续时间以秒为单位。我有一些非常相似的另一个数据源工作正常。

我尝试转换并使用第一个日期的演员，认为最后的00:00:00可能会有冲突，但无济于事。我不确定你们可能需要的更多信息，请告诉我。

* Edit1 - 我的MIN（CIC.StatusDateTime）作为Beginning，MAX（CIC.EndDateTime）作为结尾似乎没有帮助。对于代理，我有三行，它们都有不同的开始和结束时间。最小值和最大值并没有给我一整天的实际最小值和最大值。

这是源数据：

userID  StatusDateTime  StatusKey   EndDateTime StateDuration
bBarsby 9/1/2011 11:36  Out of the office   9/1/2011 11:36  0
bBarsby 9/1/2011 11:36  Out of the office   9/1/2011 11:36  5
bBarsby 9/1/2011 11:36  Available   9/1/2011 14:18  9711
bBarsby 9/1/2011 14:18  Away from desk  9/1/2011 14:39  1261
bBarsby 9/1/2011 14:39  Available   9/1/2011 19:51  18693
bBarsby 9/1/2011 19:51  Out of the office   9/1/2011 19:51  4
bBarsby 9/1/2011 19:51  Out of the office   9/1/2011 19:51  0

以下是结果

ST_Date UserID  Beginning   Ending  LoginTime   Lunch   Break   Coaching      Etc.
9/1/2011 0:00   bBarsby 9/1/2011 11:36  9/1/2011 14:18  9711    0   0   0   0
9/1/2011 0:00   bBarsby 9/1/2011 14:18  9/1/2011 14:39  1261    0   0   0   0
9/1/2011 0:00   bBarsby 9/1/2011 14:39  9/1/2011 19:51  18693   0   0   0   0

Answer 1

您按CIC.StatusDateTime和CIC.EndDateTime进行分组...只需：

GROUP BY DATEADD(dd, 0, DATEDIFF(dd, 0, CIC.StatusDateTime)), CIC.UserId

Answer 2

Aaron Bertrand有正确的想法。你绝对不应该在结束日期定义组。

以下是用该组重写的查询。

SELECT
  sub.ST_Date,
  sub.UserId,
  MIN(sub.StatusDateTime) as Beginning,
  MAX(sub.EndDateTime) as ending,
  DATEPART(hour,MAX(sub.EndDateTime) - MIN(sub.StatusDateTime)) * 3600
  + DATEPART(minute, MAX(sub.EndDateTime) - MIN(sub.StatusDateTime)) * 60
  + DATEPART(second,MAX(sub.EndDateTime) - MIN(sub.StatusDateTime)) AS Login_Time,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 5 THEN StateDuration ELSE 0 END) AS Lunch,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 3 THEN StateDuration ELSE 0 END) AS Break_,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 4 THEN StateDuration ELSE 0 END) AS Coaching,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 10 THEN StateDuration ELSE 0 END) AS SYS_Down,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 7 THEN StateDuration ELSE 0 END) AS Training,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 6 THEN StateDuration ELSE 0 END) AS Meeting,
  SUM(CASE WHEN sub.AGT_STATE_GRP = 11 THEN StateDuration ELSE 0 END) AS Default_
FROM
(
  SELECT 
    DATEADD(dd, 0, DATEDIFF(dd, 0, CIC.StatusDateTime)) AS ST_Date
    CIC.UserId,
    CIC.StatusDateTime,
    CIC.EndDateTime,
    RS.AGT_STATE_GRP,
    StateDuration
  FROM
    dbo.CIC_AgentActivityLog_TMP AS CIC
    JOIN dbo.REF_AGT_STATES AS RS
    ON CIC.StatusKey = RS.AGT_STATE_DESC
  WHERE CIC.StatusDateTime >='2011-09-01'
) sub
GROUP BY sub.ST_Date, sub.UserId
ORDER BY sub.ST_Date, sub.UserId

Answer 3

在组中使用ST_Date，因为这是表格中的列名。

GROUP BY CIC.StatusDateTime, CIC.UserId, CIC.EndDateTime
to
GROUP BY CIC.ST_DATE, CIC.UserId, CIC.EndDateTime

编辑：

group by DATEADD(dd, 0, DATEDIFF(dd, 0, CIC.StatusDateTime)),CIC.UserId, MAX(CIC.EndDateTime)

SQL选择不正确分组

3 个答案: