Question

我有一个视图，用于生成每周运行的关键任务cron的验证报告。该视图显示每周运行一次，贷方帐户总数，（我的功劳归功于SO伙伴！），OPEN帐户的总和，已关闭帐户的总和，基本内容。

mysql> select * from view_command_OPEN_CLOSED_tally limit 5;
+---------------------+-------------------------+-----------------------+---------------------+-------------------+----------------------+--------------------+
| created             | total_accounts_credited | total_amount_credited | total_open_credited | total_amount_open | total_closed_credited| total_amount_closed|
+---------------------+-------------------------+-----------------------+---------------------+-------------------+----------------------+--------------------+
| 2019-01-19 00:00:00 |                   18175 |               3173.68 |               16953 |           7063.68 |                 1222 |             110.00 |
| 2019-01-12 00:00:00 |                   18135 |               4768.43 |               17053 |           9358.43 |                 1082 |             410.00 |
| 2019-01-10 09:00:27 |                      80 |               1497.75 |                  80 |           1497.75 |                    0 |               0.00 |
| 2019-01-09 09:20:55 |                      51 |                933.50 |                  50 |            915.75 |                    1 |              17.75 |
| 2019-01-08 16:45:14 |                      10 |                187.50 |                  10 |            187.50 |                    0 |               0.00 |
+---------------------+-------------------------+-----------------------+---------------------+-------------------+----------------------+--------------------+

我认为我可以轻松地从该视图中生成第二个视图，并且可以使用子查询显示与上周百分比的差异。 IE浏览器：已创建，已计总贷方，上周差额，上周贷方，已计总贷方，上周贷方额，上周等... 我无法获得前一周的结果来匹配数据并计算百分比。结果集将支付前几行我尝试过的所有查询都围绕着基本的左联接，但是由于某些原因，“上周”的结果未在适当的位置显示。很明显，我的方法对MySQL表现出不良的态度，作为回应，它使我感到不连贯。

mysql> SELECT vtally.created,
    ->        vtally.total_accounts_credited,
    ->        vtally2.total_accounts_credited `previous`,
    ->        vtally.total_open_credited ,
    ->        vtally2.total_open_credited `previous`,
    ->        vtally.total_closed_credited,
    ->        vtally2.total_closed_credited`previous`
    -> FROM view_command_OPEN_CLOSED_tally vtally
    -> LEFT JOIN view_command_OPEN_CLOSED_tally vtally2
    ->   ON vtally.created = vtally2.created - INTERVAL 7 DAY
    -> GROUP BY
    ->   DATE(vtally.created)
    -> ORDER BY vtally.created DESC LIMIT 2;
+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+
| created             | total_accounts_credited | previous | total_open_credited | previous | total_closed_credited | previous |
+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+
| 2019-01-19 00:00:00 |                    8175 |     NULL |                6953 |     NULL |                  222 |     NULL |
| 2019-01-12 00:00:00 |                    8135 |     8175 |                7053 |     6953 |                  082 |     1222 |
+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+

上面的结果集是一个示例，我希望结果如下，每一个先前的列都将其前一周的结果保留在其左边：

（仅顶部行，不考虑底部行）

+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+
| created             | total_accounts_credited | previous | total_open_credited | previous | total_closed_credited | previous |
+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+
| 2019-01-19 00:00:00 |                    8175 |     8135 |                6953 |     7053 |                  222 |     082 |
| 2019-01-12 00:00:00 |                    8135 |     8175 |                7053 |     6953 |                  082 |     1222 |
+---------------------+-------------------------+----------+---------------------+----------+----------------------+----------+

在适当的列中具有适当的“上一个”总数时，我缺少什么？

我相信这可能是Group BY，但不知道从哪里去，似乎已经尝试了所有选择。

Answer 1

如果vtally2代表先前的数据，则表示vtally.created > vtally2.created。

因此，在联接中将减号更改为加号：vtally.created = vtally2.created + INTERVAL 7 DAY。

如果您使用窗口函数在子查询中计算列，则也可以忽略间隔为非恒定。

类似这样的东西：

WITH numbered_tally AS (
    select created, total_accounts_credited, total_open_credited, total_closed_credited,
           row_number() OVER (ORDER BY created) AS tally_number
    from view_command_OPEN_CLOSED_tally
)
SELECT  vtally.created,
        vtally.total_accounts_credited,
       vtally2.total_accounts_credited `previous`,
        vtally.total_open_credited ,
       vtally2.total_open_credited `previous`,
        vtally.total_closed_credited,
       vtally2.total_closed_credited `previous`
FROM numbered_tally  vtally
LEFT JOIN numbered_tally  vtally2
   ON vtally.tally_number = 1 + vtally2.tally_number

一个人如何生成一个视图，其中包含来自第二个视图的总和？

1 个答案: