1

如何转换工作时间 08:00:00-11:59:00;13:00:00-16:59:00; 转换为 48 位布尔格式,如

 "000000000000000011111111001111111100000000000000" 

其中每个数字指的是使用 Oracle SQL Query 的 30 分钟粒度?

4

1 回答 1

5

假设您从一个始终具有分号分隔的 from/to 时间对的字符串开始,最后一对之后有一个尾随分号;并且这些时间总是 HH24:MI:SS 秒总是零,如图所示;然后...您可以将字符串拆分为多个字符串对,代表每个从/到对:

select regexp_substr('08:00:00-11:59:00;13:00:00-16:59:00;', '(.*?)(;|-|$)', 1, (2 * level - 1), null, 1),
  regexp_substr('08:00:00-11:59:00;13:00:00-16:59:00;', '(.*?)(;|-|$)', 1, 2 * level, null, 1)
from dual
connect by level <= regexp_count('08:00:00-11:59:00;13:00:00-16:59:00;', ';')
REGEXP_S REGEXP_S
-------- --------
08:00:00 11:59:00
13:00:00 16:59:00

您可以在标称日内生成所有半小时块(选择一个不受 DST 开关影响的块):

select to_char(date '2000-01-01' + ((level - 1) / 48), 'HH24:MI":00"')
from dual
connect by level <= 48
TO_CHAR(
--------
00:00:00
00:30:00
01:00:00
01:30:00
02:00:00
...
23:00:00
23:30:00

然后使用字符串比较将它们连接在一起以查看重叠的位置(这就是时间格式很重要的原因);为简单起见,使用 CTE 提供初始字符串,然后提供前两个查询的结果:

with t1 (working_hours) as (
  select '08:00:00-11:59:00;13:00:00-16:59:00;' from dual
),
t2 (working_from, working_to) as (
  select regexp_substr(working_hours, '(.*?)(;|-|$)', 1, (2 * level - 1), null, 1),
    regexp_substr(working_hours, '(.*?)(;|-|$)', 1, 2 * level, null, 1)
  from t1
  connect by level <= regexp_count(working_hours, ';')
),
t3 (block_from) as (
  select to_char(date '2000-01-01' + ((level - 1) / 48), 'HH24:MI":00"')
  from dual
  connect by level <= 48
)
select block_from,
  case when t2.working_from is null then 0 else 1 end as flag
from t3
left join t2 on t2.working_from <= t3.block_from and t2.working_to >= t3.block_from
BLOCK_FROM  FLAG
----------  ----
00:00:00    0
00:30:00    0
...
07:30:00    0
08:00:00    1
08:30:00    1
...
11:00:00    1
11:30:00    1
12:00:00    0
12:30:00    0
13:00:00    1
13:30:00    1
...
16:00:00    1
16:30:00    1
17:00:00    0
...
23:00:00    0
23:30:00    0

然后最后将它们聚合成一个结果字符串:

with t1 (working_hours) as (
  select '08:00:00-11:59:00;13:00:00-16:59:00;' from dual
),
t2 (working_from, working_to) as (
  select regexp_substr(working_hours, '(.*?)(;|-|$)', 1, (2 * level - 1), null, 1),
    regexp_substr(working_hours, '(.*?)(;|-|$)', 1, 2 * level, null, 1)
  from t1
  connect by level <= regexp_count(working_hours, ';')
),
t3 (block_from) as (
  select to_char(date '2000-01-01' + ((level - 1) / 48), 'HH24:MI":00"')
  from dual
  connect by level <= 48
)
select listagg(case when t2.working_from is null then 0 else 1 end)
  within group (order by t3.block_from) as result
from t3
left join t2 on t2.working_from <= t3.block_from and t2.working_to >= t3.block_from
RESULT
------------------------------------------------
000000000000000011111111001111111100000000000000

db<>小提琴

如果您的初始字符串实际上来自一个表,并且您需要一次对多行进行此转换,那么连接拆分会稍微复杂一些,递归 CTE 可能更适合该部分。

只是为了好玩,这里有一个例子

于 2021-06-29T10:50:58.567 回答