当时间戳的开始时间大于时间戳的结束时间时查询失败

Question

我有一个基于CTE的查询，其中我检索两个给定时间间隔之间的小时间隔。我的查询工作如下：

获取开始和结束日期时间（比如说07-13-2011 00:21:09和07-31-2011 21:11:21）获得每小时间隔的小时总查询值（在这里是00到21，总共21小时，但这是参数化的，取决于我为输入提供的时间）。

此查询适用于第一个时间戳的小时小于第二个时间的输入 - 例如， 03 AM是第一个时间戳的时间，07 AM是第二个时间戳的时间，但是存在问题。当我想要检索输入的总计数时，如07-13-2011 22:11:43和07-25-2011 04:06:04，我遇到了问题。我需要检索以下查询的总计数：

07-13-2011 22:00:00 143 //representing the total amounts of queries 22:11:43 - 22:59:59 interval- 
07-13-2011 23:00:00 121 //representing the total amounts of queries in 23:00:00 -23:59:59 interval- 
07-14-2011 00:00:00  65 //00:00:00 - 00:59:59 interval 
07-14-2011 01:00:00  51 //01:00:00 - 01:59:59 interval... 
. 
. 
. 
07-14-2011 04:00:00 22 //query amount for 04:00:00 - 04:06:04 interval 

等等。除了我在下面写的CTE查询外，还需要做些什么？

WITH cal AS (
    SELECT generate_series('2011-02-02 00:00:00'::timestamp 
         , '2012-04-01 05:00:00'::timestamp 
         , '1 hour'::interval) AS stamp 
    ) 
, qqq AS (
    SELECT date_trunc('hour', calltime) AS stamp 
    , count(*) AS zcount 
    FROM mytable 
    WHERE calltime >= '07-13-2011 22:00:00' 
    AND calltime <='07-31-2011 04:33:21' 
    AND calltime::time >= '22:00:00' 
    AND calltime::time <= '04:33:21' 
    -- this calltime::time part obviously doesn't work due to common sense and logic 
    -- edited it to show what I try to mean 
    AND date_part('hour', calltime) >= 0 
    AND date_part('hour', calltime) <= 21 
    GROUP BY date_trunc('hour', calltime) 
    ) 
SELECT cal.stamp 
    , COALESCE (qqq.zcount, 0) AS zcount 
FROM cal 
LEFT JOIN qqq ON cal.stamp = qqq.stamp 
WHERE cal.stamp >= '07-13-2011 22:00:00' 
AND cal.stamp<='07-31-2011 04:33:21' 
AND date_part('hour', cal.stamp) >= 0 
AND date_part('hour', cal.stamp) <= 21 
ORDER BY stamp ASC; 

Answer 1

考虑这个修正版：

WITH param AS (
    SELECT '2011-07-13 22:11:43'::timestamp AS start -- supply start/stop once 
     ,'2011-07-25 04:06:04'::timestamp AS stop 
    ) 
    , cal AS (
    SELECT generate_series(date_trunc('hour', p.start) 
         ,date_trunc('hour', p.stop + interval '1h') 
         ,interval '1h') AS h 
    FROM param p 
    ) 
    , q AS (
    SELECT date_trunc('hour', calltime) AS h 
     ,count(*) AS ct 
    FROM mytable 
     ,param p 
    WHERE calltime >= p.start 
    AND calltime <= p.stop 
    -- uncomment if you actually want to exclude hours 22 & 23 (?) 
    -- AND extract('hour' FROM calltime) BETWEEN 0 AND 21 
    GROUP BY 1 
    ) 
SELECT cal.h, COALESCE(q.ct, 0) AS ct 
FROM cal 
LEFT JOIN q USING (h) 
-- uncomment if you actually want to exclude hours 22 & 23 (?) 
-- WHERE extract('hour' FROM cal.h) BETWEEN 0 AND 21 
ORDER BY 1; 

主要的变化是从生成的实际时间跨度小时的时候了。
删除了一些不需要的条件。
对时间戳使用ISO 8601格式（适用于每种语言环境）。

查找更多的上下文和链接在这个related answer。唯一的区别是：在那里它的运行计数。

Answer 2

这一条款：

AND calltime::time >= '00:21:09' AND calltime::time <= '21:11:21' 

看起来可疑，考虑到其他条款。

目前尚不清楚为什么需要它，并且如果文字时间分别为22:11:43和04:06:04作为示例不起作用，则该条款将过滤掉所有内容，以便每小时结束0次计数。