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。当我使用下面的查询使用pivot时，它会在下面显示结果，其中SendItemDateTime1不是分组...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

以下是上述查询的结果。

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

以下是我想要的......

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，第一次在这里发帖提问......

修改

这项工作对我有“Where”条款：

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来转移第一个表格中的数据 - 您只需要明确列出所有付款类型列。我假设SUM()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，重新在哪里过滤

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，如果过滤需要在不透明的列中完成，那么您可以像使用的那样使用CTE或派生表。以下是CTE和后置过滤器WHERE的预过滤器示例：

-- 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

我在哪里将WHERE子句放在具有日期事务的PIVOT中

1 个答案: