Question

我有一张表，每分钟都会记录一次数据。它看起来像：

time_stamp            wh   
2013-02-01 08:00:00  1700    
2013-02-01 08:01:00  1600    
2013-02-01 08:02:00  1800    
2013-02-01 08:03:00  1700    
2013-02-01 08:04:00  1900    
2013-02-01 08:05:00  1400    
2013-02-01 08:06:00  1500    
2013-02-01 08:07:00  1700    
2013-02-01 08:08:00  1700    
2013-02-01 08:09:00  1800    
2013-02-01 08:10:00  1800    
2013-02-01 08:11:00  1400

...等

我想每3分钟总结一次数据，每3分钟找一次平均值。输出结果如下：

time_stamp            SUM(wh)    AVG(wh)
2013-02-01 08:00:00    5100       1700
2013-02-01 08:03:00    5500       1833.3333
2013-02-01 08:06:00    4900       1633.3333
2013-02-01 08:09:00    5000       1666.6667

这似乎很容易，但我无法得到我想要的结果。这是我试过的SQL查询：

SELECT from_unixtime(ROUND(unix_timestamp(time_stamp) / 
       (60*3)) * 60 * 3) AS rounded_time,
       SUM(wh),
       AVG(wh)
FROM electrex_0
WHERE time_stamp BETWEEN '2013-02-01 08:00' AND '2013-02-01 09:00'
GROUP BY rounded_time

通过使用此sql代码，我设法正确获取time_stamp，但不是SUM和AVG。我意识到它将以另一种方式总结。例如：2013-02-01 08:03:00的Sum(wh)是5100，它将从08：02,08：03,08：04获取数据并总结一下。

Answer 1

你想如何从5500获得：

2013-02-01 08:03:00  1700    
2013-02-01 08:04:00  1900    
2013-02-01 08:05:00  1400

总计达5000，不是吗？

以下是您的查询已修复以返回正确的数据：

SELECT
  from_unixtime(ROUND((unix_timestamp(time_stamp) - 60) / (60*3)) * 60 * 3) AS rounded_time,
  SUM(wh),
  AVG(wh)
FROM electrex_0
WHERE time_stamp BETWEEN '2013-02-01 08:00'
    AND '2013-02-01 09:00'
GROUP BY rounded_time

返回：

ROUNDED_TIME                    SUM(WH) AVG(WH)
February, 01 2013 08:00:00+0000 5100    1700
February, 01 2013 08:03:00+0000 5000    1666.6667
February, 01 2013 08:06:00+0000 4900    1633.3333
February, 01 2013 08:09:00+0000 5000    1666.6667

Demo

Answer 2

以下是使用变量

的方法

SET @VAR:=1;
SET @group=0;
SELECT 
    time_stamp,
    SUM(wh) AS `TotalSum`,
    AVG(wh) AS `TotalAvg`
FROM 
  (
    SELECT 
        time_stamp,
        wh,
        @VAR AS Level,
        @group AS `TGroup`,
        @group := IF(@VAR = 3,@group + 1 ,@group) ,
        @VAR := IF(@VAR = 3,1,@VAR+1 )
    FROM electrex_0
  ) AS l
GROUP BY TGroup

SQL Fiddle Demo

输出

|     TIME_STAMP                  | TOTALSUM |  TOTALAVG |
----------------------------------------------------------
| February, 01 2013 08:00:00+0000 |     5100 |      1700 |
| February, 01 2013 08:03:00+0000 |     5000 | 1666.6667 |
| February, 01 2013 08:06:00+0000 |     4900 | 1633.3333 |
| February, 01 2013 08:09:00+0000 |     5000 | 1666.6667 |

将3分钟记录合并并总结为单行记录

2 个答案:

SQL Fiddle Demo