PSQL：隨着日期範圍的表，我該怎麼計算的，與所選擇的日期範圍

Question

我有一個看起來像這樣的日期範圍的表重疊時間量，我給他們打電話Attempts：

id |   start   |   end 
-----+----------------------------+---------------------------- 
1 | 2017-09-09 01:09:00  | 2017-09-09 02:09:00 
58 | 2017-09-09 08:35:51.420588 | 2017-09-09 17:01:57.106442 
59 | 2017-09-09 17:12:13.644954 | 2017-09-09 17:17:42.186903 
60 | 2017-09-09 17:18:51.794184 | 2017-09-09 17:50:09.508929 
61 | 2017-09-09 17:19:31.205516 | 2017-09-09 17:20:11.924069 
62 | 2017-09-09 17:20:32.320789 | 2017-09-09 17:23:19.973078 
63 | 2017-09-09 17:23:53.849006 | 2017-09-09 17:24:23.574369 
64 | 2017-09-09 17:24:56.333346 | 2017-09-09 17:27:10.556657 
65 | 2017-09-09 17:27:12.605315 | 2017-09-09 17:27:14.680357 

我想做一個查詢，將返回該表和提供的日期範圍內的重疊時間總量。例如，我想在9月份獲得嘗試時間的總數。

Answer 1

在input_range你把參數。

首先，你需要的，如果一排重疊檢查您的範圍

WHERE "start" < range_end 
    AND "end" > range_start 

但是，如果該行超出該範圍重新分配範圍的邊界，而不是

CASE WHEN "start" < range_start 
    THEN range_start 
    ELSE "start" 
END as new_start 

再總結秒beteeen時間戳

SUM (EXTRACT(EPOCH FROM (new_end - new_start))) 

SQL DEMO

WITH input_range as (
    SELECT '2017-09-09 17:18'::timestamp as range_start, 
      '2017-09-09 17:22'::timestamp as range_end 
), fix_data as (
    SELECT A.*, CASE WHEN "start" < range_start 
        THEN range_start 
        ELSE "start" 
       END as new_start 
       , CASE WHEN "end" > range_end 
        THEN range_end 
        ELSE "end" 
       END as new_end  
    FROM Attempts A 
    CROSS JOIN input_range 
    WHERE "start" < range_end 
     AND "end" > range_start 
) 
SELECT SUM (EXTRACT(EPOCH FROM (new_end - new_start))) 
     -- * to see what are you SUM 
FROM fix_data;