我有一张如下表格。
+------------+------------------+-------------------------------+-------------+
| day | workflow_step_id | UNIX_TIMESTAMP(finished_time) | workflow_id |
+------------+------------------+-------------------------------+-------------+
| 2014-04-30 | 1 | 1398852780 | 1 |
| 2014-04-30 | 17 | 1398871213 | 2 |
| 2014-04-30 | 6 | 1398872807 | 1 |
| 2014-04-30 | 22 | 1398898676 | 2 |
| 2014-04-30 | 11 | 1398898234 | 1 |
| 2014-04-30 | 16 | 1398866813 | 2 |
| 2014-04-30 | 5 | 1398869940 | 1 |
| 2014-04-30 | 21 | 1398893419 | 2 |
| 2014-04-30 | 10 | 1398894136 | 1 |
| 2014-04-30 | 15 | 1398861385 | 2 |
| 2014-04-30 | 4 | 1398860271 | 1 |
| 2014-04-30 | 20 | 1398888252 | 2 |
| 2014-04-30 | 9 | 1398886916 | 1 |
| 2014-04-30 | 14 | 1398863922 | 2 |
| 2014-04-30 | 3 | 1398865682 | 1 |
| 2014-04-30 | 19 | 1398881994 | 2 |
| 2014-04-30 | 8 | 1398882497 | 1 |
| 2014-04-30 | 13 | 1398852785 | 2 |
| 2014-04-30 | 2 | 1398856674 | 1 |
| 2014-04-30 | 18 | 1398878836 | 2 |
| 2014-04-30 | 7 | 1398878949 | 1 |
| 2014-04-30 | 12 | 1398850920 | 2 |
| 2014-05-01 | 12 | 1398932040 | 2 |
| 2014-05-01 | 1 | 1398938880 | 1 |
| 2014-05-01 | 17 | 1398957830 | 2 |
| 2014-05-01 | 6 | 1398961385 | 1 |
| 2014-05-01 | 22 | 1398996388 | 2 |
| 2014-05-01 | 11 | 1398994543 | 1 |
| 2014-05-01 | 16 | 1398946714 | 2 |
| 2014-05-01 | 5 | 1398959182 | 1 |
| 2014-05-01 | 21 | 1398989773 | 2 |
| 2014-05-01 | 10 | 1398979568 | 1 |
| 2014-05-01 | 15 | 1398949939 | 2 |
| 2014-05-01 | 4 | 1398953732 | 1 |
| 2014-05-01 | 20 | 1398980246 | 2 |
| 2014-05-01 | 9 | 1398971281 | 1 |
| 2014-05-01 | 14 | 1398940775 | 2 |
| 2014-05-01 | 3 | 1398944207 | 1 |
| 2014-05-01 | 19 | 1398970295 | 2 |
| 2014-05-01 | 8 | 1398967778 | 1 |
| 2014-05-01 | 13 | 1398940935 | 2 |
| 2014-05-01 | 2 | 1398947433 | 1 |
| 2014-05-01 | 18 | 1398966959 | 2 |
| 2014-05-01 | 7 | 1398965931 | 1 |
-------------------------------------------------------------------------------
命令：
SELECT day,
stats.workflow_step_id,
Max(Unix_timestamp(finished_time)) - Min(Unix_timestamp(finished_time)),
workflow_id
FROM modeling_dashboard_workflow_stats stats
INNER JOIN modeling_dashboard_workflow_step step
ON stats.workflow_step_id = step.workflow_step_id
ORDER BY day;
在此示例中，只有两个day和两个workflow_id，但它可以有day个或workflow_id个。
我想计算每天的每个workflow_id的最大时间戳和最小时间戳之间的差异。
看起来应该是这样的。
+------------+-------------------------------------------------------------+
| day | MAX(timestamp) - MIN(timestamp) |
+------------+-------------------------------------------------------------+
| 2014-04-30 | difference bw max&min timestamp with workflow_id=1 in 04-30 |
| 2014-04-30 | difference bw max&min timestamp with workflow_id=2 in 04-30 |
| 2014-05-01 | difference bw max&min timestamp with workflow_id=1 in 05-01 |
| 2014-05-01 | difference bw max&min timestamp with workflow_id=2 in 05-01 |
----------------------------------------------------------------------------
我该怎么做？
理想情况下，它会将workflow_id = 1的结果和workflow_id = 2的结果放在不同的列中（日期，MAX = MIN表示id = 1，MAX-MIN表示id = 2），但这只是我的结果下一步。
答案 0 :(得分：1)
试试这个：
select day
,workflow_step_id
,MAX(UNIX_TIMESTAMP(finished_time)) - MIN(UNIX_TIMESTAMP(finished_time))
from modeling_dashboard_workflow_stats
group by day
,workflow_step_id
答案 1 :(得分：1)
SELECT day, workflow_id
, MAX(UNIX_TIMESTAMP(finished_time)) - MAX(UNIX_TIMESTAMP(finished_time))
FROM modeling_dashboard_workflow_stats
GROUP BY day, workflow_id