c# - 我应该在带有日期事务的 PIVOT 中的何处放置 WHERE 子句

Question

amount_usd  paytype SendItemDateTime1
5.00    google  2015-04-01
2.00    google  2015-04-01
5.00    transfer    2015-04-01
15.00   google  2015-04-01
5.00    google  2015-04-01
2.00    google  2015-04-02
60.00   google  2015-04-02
60.00   google  2015-04-02
5.00    google  2015-04-03

上面是我的演示数据库，其中包含 amount_usd、paytype 和 SendItemDateTime1 列。当我在下面的查询中使用 pivok 时，结果如下，其中 SendItemDateTime1 不是分组依据...有什么问题？

 select amount_usd, paytype, SendItemDateTime1 from tblMobile_RequestOrderLog
  where status = 1 and sendstatus = 1 and enable = 1
  and SendItemDateTime1  between '4/1/2015' and '4/30/2015'
  order by SendItemDateTime1 

下面是上面查询的结果。

SenditemDateTime1   google  mol molpay  molstore    paypal  transfer
2015-04-01  15.00   NULL    NULL    NULL    NULL    NULL
2015-04-01  5.00    NULL    NULL    NULL    NULL    NULL
2015-04-01  15.00   NULL    NULL    NULL    NULL    NULL
2015-04-01  5.00    NULL    NULL    NULL    NULL    NULL
2015-04-01  60.00   NULL    NULL    NULL    NULL    NULL
2015-04-01  10.00   NULL    NULL    NULL    NULL    NULL

下面是我想要的...

SendItemDate    google  mol molpay  molstore    paypal  transfer
2015-04-01      32      0   0       0          0    5
2015-04-02      122     0   0       0          0    0
2015-04-03      5       0   0       0          0    0

抱歉，第一次在这里发帖...

编辑

这对我来说有“Where”子句:

SELECT SendItemDateTime1, COALESCE([google], 0), COALESCE([transfer], 0),
       COALESCE([paypal], 0),COALESCE([molpay], 0)
FROM (Select SendItemDateTime1, paytype, amount_usd 
      from tblMobile_RequestOrderLog
      where gameidn = 248 and status = 1 and sendstatus = 1 and enable = 1 
            and SendItemDateTime1 between '4/1/2015 12:00:00 AM'
                                      and '4/30/2015 11:59:59'
) X 
PIVOT
(
  SUM(amount_usd)
  for [paytype] IN ([google],[transfer],[paypal],[molpay])
) piv;

Answer 1

您可以使用以下查询对第一个表中的数据进行透视 - 您只需要明确列出所有付款类型列。我假设 SUM() 作为要应用的聚合:

SELECT SendItemDateTime1, [google],[transfer],[paypal],[molpay]
FROM MyTable
PIVOT
(
 SUM(amount_usd)
 for [paytype] IN ([google],[transfer],[paypal],[molpay])
) piv;

SqlFiddle here

编辑，我在哪里过滤

如果过滤谓词可以应用于最终列，则可以在 PIVOT 之后应用 WHERE。否则，如果过滤需要在非透视列中完成，那么您可以使用 CTE 或派生表，就像您所做的那样。这是 CTE 中前置过滤器和后置过滤器 WHERE 的示例:

-- Prefilter of non-pivot columns done in CTE or Derived table
WITH cte AS
(
  SELECT [amount_usd], [paytype], [SendItemDateTime1]
  FROM MyTable
  WHERE [amount_usd] > 2
)
SELECT SendItemDateTime1, COALESCE([google], 0), COALESCE([transfer], 0),
       COALESCE([paypal], 0),COALESCE([molpay], 0)
FROM cte
PIVOT
(
 SUM(amount_usd)
 for [paytype] IN ([google],[transfer],[paypal],[molpay])
) piv
-- Post filter of pivot columns done on the final projection
WHERE SendItemDateTime1 > '2015-01-01';

Updated Fiddle