如果此查询返回的日期在请求的范围内。
select created_at from user where created_at between '2015-01-06 00:00:00.000000' and '2015-03-06 00:00:00.000000'
有没有一种方法可以得到这样的结果,但是有多个日期包含在一个日期数组中。
我只是想举一个例子。我有这个日期数组,总是会有第一个和最后一个日期。
Array['2015-01-06 00:00:00.000000','2015-02-10 15:17:18.895000' <- First range
'2017-10-05 14:41:04.191000','2017-10-11 14:49:36.454000' <- Second range
有没有办法放置这样的脚本?
select created_at from win_users
where (created_at between [First Date] and [Second Date])
or (created_at between [Third Date] and [Fourth Date])
但是不使用循环来连接where语句吗?
答案 0 :(得分:1)
在这种情况下,日期数组非常不舒服。使用daterange
和containtment operator <@
的数组,例如:
with my_table(id, created_at) as (
values
(1, '2015-01-10'::timestamp),
(2, '2016-05-10'),
(3, '2017-10-10')
)
select *
from my_table
where created_at::date <@ any(array[daterange('2015-01-06','2015-02-10'), daterange('2017-10-05','2017-10-11')])
id | created_at
----+---------------------
1 | 2015-01-10 00:00:00
3 | 2017-10-10 00:00:00
(2 rows)
如果您绝对要使用日期数组(坦白地说,我不这么认为),请使用此函数将其转换为daterange数组:
create or replace function date_pairs_to_ranges(date[])
returns daterange[] language sql as $$
select array_agg(daterange(d1, d2))
from unnest($1) with ordinality as u1(d1, o1)
join unnest($1) with ordinality as u2(d2, o2)
on o1/ 2* 2 < o1 and o2 = o1+ 1
$$;
with my_table(id, created_at) as (
values
(1, '2015-01-10'::timestamp),
(2, '2016-05-10'),
(3, '2017-10-10')
)
select *
from my_table
where created_at::date <@ any(date_pairs_to_ranges(array['2015-01-06','2015-02-10','2017-10-05','2017-10-11']::date[]))
答案 1 :(得分:0)
如果范围数组始终采用您发布的格式(即4个项目,前两个元素是第一个范围,后两个元素是第二个范围),则可以编写如下查询:
WITH ranges AS (
SELECT '{2015-01-06 00:00:00.000000,2015-02-10 15:17:18.895000,2017-10-05 14:41:04.191000,2017-10-11 14:49:36.454000}'::date[] dates
)
SELECT win_users.created_at FROM win_users, ranges
WHERE (win_users.created_at > ranges.dates[1] AND win_users.created_at < ranges.dates[2]) OR (win_users.created_at > ranges.dates[3] AND win_users.created_at < ranges.dates[4]);