需要每一列的总数，并为空数据点显示 0

Question

我有一个脚本（如下），它有多个连接的列，我需要在底部的最后一行包含每列的总和。我知道 sum 函数，但我不知道将它放在脚本中的什么位置。我还想知道如何或在何处合并此语句 SELECT ISNULL(myColumn, 0 ) FROM myTable 因此空值显示 0.

SELECT A.[start_tran_hour],
       [singles picked],
       [single packed],
       [multis picked],
       [units sorted],
       [multis packed],
       [sa packed],
       [sa picked],
       [total picked],
       [total packed],
       [total shipped]
FROM   (SELECT Datepart(hour, start_tran_time)start_tran_hour,
               Sum(tran_qty)                  AS 'Total Shipped'
        FROM   t_tran_log WITH(nolock)
        WHERE  tran_type IN ( '340', '341' )
               AND Cast(start_tran_date AS DATE) = '2021-07-06'
        GROUP  BY Datepart(hour, start_tran_time))A
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Total Picked'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '301'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))B
              ON A.start_tran_hour = B.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Single Packed'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '315'
                         AND description = 'Single Packing'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))C
              ON A.start_tran_hour = C.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Singles Picked'
                  FROM   t_tran_log WITH(nolock)
                         LEFT JOIN t_order WITH(nolock)
                                ON
                         t_tran_log.control_number = t_order.order_number
                  WHERE  tran_type = '301'
                         AND t_order.route = 'SINGLE'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))D
              ON A.start_tran_hour = D.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Multis Picked'
                  FROM   t_tran_log WITH(nolock)
                         LEFT JOIN t_order WITH(nolock)
                                ON
                         t_tran_log.control_number = t_order.order_number
                  WHERE  tran_type = '301'
                         AND t_order.route = 'MULTI'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))E
              ON A.start_tran_hour = E.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Multis Packed'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '315'
                         AND description = 'Multi Packing'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))F
              ON A.start_tran_hour = F.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'SA Picked'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '301'
                         AND ( location_id LIKE 'PR%'
                                OR location_id LIKE 'SA%' )
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))G
              ON A.start_tran_hour = G.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'SA Packed'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '315'
                         AND ( location_id LIKE 'PACKSA%' )
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))H
              ON A.start_tran_hour = H.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Units Sorted'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '311'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY Datepart(hour, start_tran_time))I
              ON A.start_tran_hour = I.start_tran_hour
       LEFT JOIN (SELECT Datepart(hour, start_tran_time)start_tran_hour,
                         Sum(tran_qty)                  AS 'Total Packed'
                  FROM   t_tran_log WITH(nolock)
                  WHERE  tran_type = '315'
                         AND Cast(start_tran_date AS DATE) = '2021-07-06'
                  GROUP  BY ( Datepart(hour, start_tran_time) ))J
              ON A.start_tran_hour = J.start_tran_hour
ORDER  BY A.start_tran_hour

Answer 1

对于延误，我们深表歉意。我一直被 Picked and Packed 绊倒。

也许这会让您开始减少条件聚合

您可能注意到我注释掉了 [singles picked] 和 [multis picked]。看起来需要 JOIN，但我对您的数据了解不足，无法做出假设。

Select start_tran_hour = Datepart(hour, start_tran_time)
      --,[singles picked]  
      ,[single packed] = sum( case when tran_type in ('315') 
                                    and description = 'Single Packing'      then tran_qty else 0 end)
      --,[multis picked]
      ,[units sorted]  = sum( case when tran_type in ('311')                then tran_qty else 0 end)
      ,[multis packed] = sum( case when tran_type in ('315') 
                                    and description='Multi Packing'         then tran_qty else 0 end)
      ,[sa packed]     = sum( case when tran_type in ('315') 
                                    and location_id LIKE 'PACKSA%'          then tran_qty else 0 end)
      ,[sa picked]     = sum( case when tran_type in ('301') 
                                    and left(location_id,2) in ('PR','SA')  then tran_qty else 0 end)
      ,[Total Picked]  = sum( case when tran_type in ('301')                then tran_qty else 0 end)
      ,[Total Packed]  = sum( case when tran_type in ('315')                then tran_qty else 0 end)
      ,[Total Shipped] = sum( case when tran_type in ('340','341')          then tran_qty else 0 end)
 From  t_tran_log
 Where Cast(start_tran_date AS DATE) = '2021-07-06'
 Group By Grouping Sets ( 
                         (Datepart(hour, start_tran_time))
                        ,()
                        )

Answer 2

我假设您有几个小时，列中的每个度量值都为空（或者根本没有），所以 SUM(all_null_values) = NULL 而您希望它为 0。

这将处理空值并添加总计行。

SELECT A.[start_tran_hour],
, sum(coalesce([singles picked], 0)) as 'singles picked'
, sum(coalesce([single packed],  0)) as 'single packed'
, sum(coalesce([multis picked],  0)) as 'multis picked'
, sum(coalesce([units sorted],   0)) as 'units sorted'
, sum(coalesce([multis packed],  0)) as 'multis packed'
, sum(coalesce([sa packed],      0)) as 'sa packed'
, sum(coalesce([sa picked],      0)) as 'sa picked'
, sum(coalesce([total picked],   0)) as 'total picked'
, sum(coalesce([total packed],   0)) as 'total packed'
, sum(coalesce([total shipped],  0)) as 'total shipped'

FROM   (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Total Shipped'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type IN ( '340', '341' )
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)A
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Total Picked'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '301'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)B ON A.start_tran_hour = B.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Single Packed'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '315'
    AND description = 'Single Packing'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)C ON A.start_tran_hour = C.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Singles Picked'
  FROM   t_tran_log WITH(nolock)
    LEFT JOIN t_order WITH(nolock) ON t_tran_log.control_number = t_order.order_number
  WHERE  tran_type = '301'
    AND t_order.route = 'SINGLE'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)D ON A.start_tran_hour = D.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Multis Picked'
  FROM   t_tran_log WITH(nolock)
    LEFT JOIN t_order WITH(nolock) ON t_tran_log.control_number = t_order.order_number
  WHERE  tran_type = '301'
    AND t_order.route = 'MULTI'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)E ON A.start_tran_hour = E.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Multis Packed'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '315'
    AND description = 'Multi Packing'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
)F ON A.start_tran_hour = F.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'SA Picked'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '301'
    AND ( location_id LIKE 'PR%'
       OR location_id LIKE 'SA%' )
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
) G ON A.start_tran_hour = G.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'SA Packed'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '315'
    AND ( location_id LIKE 'PACKSA%' )
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
) H ON A.start_tran_hour = H.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Units Sorted'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '311'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY Datepart(hour, start_tran_time)
) I ON A.start_tran_hour = I.start_tran_hour
LEFT JOIN (
  SELECT Datepart(hour, start_tran_time) as start_tran_hour
  , Sum(tran_qty) AS 'Total Packed'
  FROM   t_tran_log WITH(nolock)
  WHERE  tran_type = '315'
    AND Cast(start_tran_date AS DATE) = '2021-07-06'
  GROUP  BY ( Datepart(hour, start_tran_time) )
) J ON A.start_tran_hour = J.start_tran_hour

group by rollup (A.start_tran_hour)

ORDER  BY A.start_tran_hour

但是，它仍然包含来自您的原始查询的 LEFT INNER JOIN（不是真实的东西。我编造了这个词。）。

FROM  t_tran_log
  LEFT JOIN t_order ON t_tran_log.control_number = t_order.order_number
WHERE tran_type = '301'
    AND t_order.route = 'MULTI'

在 WHERE 子句中包含 t_order.route = 'MULTI' 会导致 LEFT JOIN 成为 INNER JOIN。所以我们要解决这个问题。

不过约翰·卡佩莱蒂的建议也不错。让我们尝试在 FROM 子句中引用一次 table 并计算 SELECT 子句中的每一列。

SELECT Datepart(hour, start_tran_time) as start_tran_hour,
, sum(
  case
    when tran_type = '301'
      AND t_order.route = 'SINGLE'
      then [singles picked]
    else 0
  end) as 'singles picked'
, sum(
  case
    when tran_type = '315'
      AND description = 'Single Packing'
      then [single packed]
    else 0
  end) as 'single packed'
, sum(
  case
    when tran_type = '301'
      AND t_order.route = 'MULTI'
      then [multis picked]
    else 0
  end) as 'multis picked'
, sum(
  case
    when tran_type = '311'
      then [units sorted]
    else 0
  end) as 'units sorted'
, sum(
  case
    when tran_type = '315'
      AND description = 'Multi Packing'
      then [multis packed]
    else 0
  end) as 'multis packed'
, sum(
  case
    when tran_type = '315'
      AND location_id LIKE 'PACKSA%'
      then [sa packed]
    else 0
  end) as 'sa packed'
, sum(
  case
    when tran_type = '301'
      AND ( location_id LIKE 'PR%'
         OR location_id LIKE 'SA%' )
      then [sa picked]
    else 0
  end) as 'sa picked'
, sum(
  case
    when tran_type = '301'
      then [total picked]
    else 0
  end) as 'total picked'
, sum(
  case
    when tran_type = '315'
      then [total packed]
    else 0
  end) as 'total packed'
, sum(
  case
    when tran_type IN ( '340', '341' )
      then [total shipped]
    else 0
  end) as 'total shipped'

from t_tran_log WITH(nolock)
  LEFT JOIN t_order WITH(nolock) ON t_tran_log.control_number = t_order.order_number
                                AND t_order.route in ('SINGLE', 'MULTI')

where Cast(start_tran_date AS DATE) = '2021-07-06'
  and tran_type IN ('340', '341', '301', '315', '311')

group by rollup (Datepart(hour, start_tran_time))

ORDER BY Datepart(hour, start_tran_time)

如果没有数据库架构和示例数据，我无法确定其中任何一项是否有效。下次，创建并引用一个 dbfiddle。

需要每一列的总数，并为空数据点显示 0

Need a total for each column and to show 0 for null data points

sql

tsql

sql-server

sql-server-2016