我目前有一个表格，显示特定productionID及其runId在特定时间计算的项目数（生产可以在不同的机器上运行）。见下文。
根据这些数据，我希望在相同的tmTimestamps中计算整个生产。 如果我只是按生产ID和tmTimestamp分组，我得到正确的值，直到最后一个时间戳，其中我只有一个ProductionRunID的值。
编辑：示例：我们在2014-04-08 22：17：38.000只有一个来自productionRunID 3080的值，这是37501.问题是我们也从其他productionRunID 3079生成了19802，但这行已经停止报告，因为生产运行在那里结束。
有什么建议吗？
我有什么：
+--------------+-----------------+-------------------------+---------+
| ProductionID | ProductionRunID | tmTimeStamp | Counted |
+--------------+-----------------+-------------------------+---------+
| 889929 | 3079 | 2014-04-08 20:38:01.000 | 0 |
| 889929 | 3079 | 2014-04-08 20:43:01.000 | 0 |
| 889929 | 3079 | 2014-04-08 20:48:02.000 | 2756 |
| 889929 | 3079 | 2014-04-08 20:53:02.000 | 8746 |
| 889929 | 3079 | 2014-04-08 20:58:03.000 | 15005 |
| 889929 | 3079 | 2014-04-08 21:03:04.000 | 19477 |
| 889929 | 3079 | 2014-04-08 21:04:10.000 | 19802 |
| 889929 | 3079 | 2014-04-08 21:09:11.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:09:11.000 | 0 |
| 889929 | 3080 | 2014-04-08 21:14:11.000 | 0 |
| 889929 | 3079 | 2014-04-08 21:14:11.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:19:12.000 | 5216 |
| 889929 | 3079 | 2014-04-08 21:19:12.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:24:13.000 | 11480 |
| 889929 | 3079 | 2014-04-08 21:24:13.000 | 19802 |
| 889929 | 3079 | 2014-04-08 21:29:14.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:29:14.000 | 17791 |
| 889929 | 3080 | 2014-04-08 21:34:15.000 | 24065 |
| 889929 | 3079 | 2014-04-08 21:34:15.000 | 19802 |
| 889929 | 3079 | 2014-04-08 21:39:16.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:39:16.000 | 30356 |
| 889929 | 3080 | 2014-04-08 21:43:03.000 | 34881 |
| 889929 | 3079 | 2014-04-08 21:43:03.000 | 19802 |
| 889929 | 3079 | 2014-04-08 21:48:04.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:48:04.000 | 37501 |
| 889929 | 3080 | 2014-04-08 21:50:44.000 | 37501 |
| 889929 | 3079 | 2014-04-08 21:50:44.000 | 19802 |
| 889929 | 3079 | 2014-04-08 21:51:54.000 | 19802 |
| 889929 | 3080 | 2014-04-08 21:51:54.000 | 37501 |
| 889929 | 3080 | 2014-04-08 21:56:55.000 | 37501 |
| 889929 | 3079 | 2014-04-08 21:56:55.000 | 19802 |
| 889929 | 3079 | 2014-04-08 22:01:56.000 | 19802 |
| 889929 | 3080 | 2014-04-08 22:01:56.000 | 37501 |
| 889929 | 3079 | 2014-04-08 22:02:37.000 | 19802 |
| 889929 | 3080 | 2014-04-08 22:02:37.000 | 37501 |
| 889929 | 3080 | 2014-04-08 22:07:37.000 | 37501 |
| 889929 | 3080 | 2014-04-08 22:12:37.000 | 37501 |
| 889929 | 3080 | 2014-04-08 22:17:38.000 | 37501 |
| 889929 | 3080 | 2014-04-08 22:22:38.000 | 37501 |
+--------------+-----------------+-------------------------+---------+
我想要的是什么：
+--------------+-------------------------+---------+
| ProductionID | tmTimeStamp | Counted |
+--------------+-------------------------+---------+
| 889929 | 2014-04-08 20:17:58.000 | 0 |
| 889929 | 2014-04-08 20:22:59.000 | 0 |
| 889929 | 2014-04-08 20:27:59.000 | 0 |
| 889929 | 2014-04-08 20:33:00.000 | 0 |
| 889929 | 2014-04-08 20:38:01.000 | 0 |
| 889929 | 2014-04-08 20:43:01.000 | 0 |
| 889929 | 2014-04-08 20:48:02.000 | 2756 |
| 889929 | 2014-04-08 20:53:02.000 | 8746 |
| 889929 | 2014-04-08 20:58:03.000 | 15005 |
| 889929 | 2014-04-08 21:03:04.000 | 19477 |
| 889929 | 2014-04-08 21:04:10.000 | 19802 |
| 889929 | 2014-04-08 21:09:11.000 | 19802 |
| 889929 | 2014-04-08 21:14:11.000 | 19802 |
| 889929 | 2014-04-08 21:19:12.000 | 25018 |
| 889929 | 2014-04-08 21:24:13.000 | 31282 |
| 889929 | 2014-04-08 21:29:14.000 | 37593 |
| 889929 | 2014-04-08 21:34:15.000 | 43867 |
| 889929 | 2014-04-08 21:39:16.000 | 50158 |
| 889929 | 2014-04-08 21:43:03.000 | 54683 |
| 889929 | 2014-04-08 21:48:04.000 | 57303 |
| 889929 | 2014-04-08 21:50:44.000 | 57303 |
| 889929 | 2014-04-08 21:51:54.000 | 57303 |
| 889929 | 2014-04-08 21:56:55.000 | 57303 |
| 889929 | 2014-04-08 22:01:56.000 | 57303 |
| 889929 | 2014-04-08 22:02:37.000 | 57303 |
| 889929 | 2014-04-08 22:07:37.000 | 57303 |
| 889929 | 2014-04-08 22:12:37.000 | 57303 |
| 889929 | 2014-04-08 22:17:38.000 | 57303 |
| 889929 | 2014-04-08 22:22:38.000 | 57303 |
| 889929 | 2014-04-08 22:27:39.000 | 57303 |
| 889929 | 2014-04-08 22:32:39.000 | 57303 |
| 889929 | 2014-04-08 22:37:40.000 | 57303 |
| 889929 | 2014-04-08 22:42:41.000 | 57303 |
| 889929 | 2014-04-08 22:47:42.000 | 57303 |
| 889929 | 2014-04-08 22:52:43.000 | 57303 |
| 889929 | 2014-04-08 22:57:44.000 | 57303 |
| 889929 | 2014-04-08 22:59:26.000 | 57303 |
| 889929 | 2014-04-08 23:04:27.000 | 57303 |
| 889929 | 2014-04-08 23:04:47.000 | 57303 |
+--------------+-------------------------+---------+
EDIT2：添加了示例结果。
答案 0 :(得分：1)
您的问题仍然很难说明，这使得其他人难以提供帮助。我从您的各种评论中收集到您希望结果包含counted列的小计，该列是同一counted的所有按时间顺序排列的行中productionId的总和productionRunId。这个的SQL是：
select t1.ProductionID, t1.ProductionRunID, t1.tmTimeStamp, t1.counted,
(
select sum(counted)
from prod as t2
where t1.ProductionID = t2.ProductionID
and t1.ProductionRunID = t2.ProductionRunID
and t1.tmTimeStamp >= t2.tmTimeStamp
) as sub_total
from prod as t1
order by 1, 2, 3
;
您尚未说明您正在使用哪个数据库，而更复杂的SQL通常会针对特定于实现（或者使用特定数据库的功能更简单）。我假设了mySQL。
这是一个小提琴来说明它：http://sqlfiddle.com/#!2/5f794/7
请注意，在您的原始问题中，您声明您想要group by时间戳 - 这是错误的并且导致读者误解了问题。请在下次尝试更清楚，并举例说明所需的结果和表格结构。
答案 1 :(得分：0)
我认为这就是你要找的东西：
select
ProductionId
,ProductionRunId
,tmTimeStamp
,sum(Counted)
from
ProdTable-or-whatever-it-is-called
group by
ProductionId
,ProductionRunId
,tmTimeStamp
order by
ProductionId
,ProductionRunId
,tmTimeStamp