使用与postgres不同的日期和聚合函数

Question

我想要每个月获得一组平均值.我正在使用DATE_TRUNC将日期时间转换为月份,但我仍然无法使用不同的DATE.似乎完全忽略它.

我只需要6行,每个月在RANGE中一​​个,但我得到1000.

SELECT DISTINCT(date_trunc('month', event_logs.start_at)) AS start_at, 
AVG(event_logs.head_count) as head_count FROM "event_logs" 
WHERE ("event_logs"."state" IN ('completed')) AND ("event_logs"."start_at" 
BETWEEN '2013-05-09 10:12:58.824846' AND '2013-11-09') 
GROUP BY start_at

还尝试投掷DISTINCT(DATE(date_trunc())但是没有做任何事情吗？DATES有什么特别之处我还没有看到吗？

+----+---------------------+---------------------------+
| id | head_count          | start_at                  |
+----+---------------------+---------------------------+
|    | 17.0                | 2013-06-01 01:00:00 +0100 |#WHY???!?!?!
|    | 15.0                | 2013-06-01 01:00:00 +0100 |#YOU ARE CLONES!
|    | 40.5                | 2013-06-01 01:00:00 +0100 |#NOT DISTINCT!
|    | 32.5                | 2013-10-01 01:00:00 +0100 |
|    | 69.0                | 2013-08-01 01:00:00 +0100 |
|    | 34.9                | 2013-10-01 01:00:00 +0100 |
|    | 9.0                 | 2013-07-01 01:00:00 +0100 |

Answer 1

试试这个:

SELECT
date_trunc('month', event_logs.start_at) AS start_at, 
AVG(event_logs.head_count) as head_count 
FROM "event_logs" 
WHERE ("event_logs"."state" IN ('completed')) 
AND ("event_logs"."start_at" 
BETWEEN '2013-05-09 10:12:58.824846' AND '2013-11-09') 
GROUP BY date_trunc('month', event_logs.start_at)

您的问题出在GROUP BY语句中.您需要GROUP BY月而不是日期/时间值.而且您不需要将DISTINCT与GROUP BY一起使用