Question

我正在创建一个选择，它使我超出了表格的结果，范围为15分钟。

我要做的是将小时从00:00到00:15分组在0:15范围内。在0:30范围内将寄存器在0:16和0:30之间分组。我将在一天中的所有时间做这件事。

下面，我描述我的选择。如果有人可以帮助我发展思路。 Tks很多。

CREATE OR REPLACE FORCE VIEW "ARADMIN"."GSC_VW_INC_DIARIOS"("ID_INCIDENTE", "STATUS_INCIDENTE", "DATAHORA_CRIACAO", "DATA_CRIACAO",       "HORA_CRIACAO", "PRIORIDADE", "IMPACTO", "URGÊNCIA")
AS
 SELECT T2318.C1,
CASE (T2318.C7)
  WHEN 0
  THEN 'NOVO'
  WHEN 1
  THEN 'DESIGNADO'
  WHEN 2
  THEN 'EM ANDAMENTO'
  WHEN 3
  THEN 'PENDENTE'
  WHEN 4
  THEN 'RESOLVIDO'
  WHEN 5
  THEN 'FECHADO'
  WHEN 6
  THEN 'CANCELADO'
END,
TO_CHAR(secs_to_datetime(T2318.C3),'DD/MM/YYYY HH24:MI:SS'),
TO_CHAR(secs_to_date(T2318.C3),'DD/MM/YYYY'),
CASE TO_CHAR(secs_to_hour(T2318.C3),'HH24:MI')
  WHEN TO_CHAR(secs_to_hour(T2318.C3),'HH24:MI') BETWEEN ('00:00' AND '00:15')
  THEN '00:15'
END,
CASE (T2318.C1000000164)
  WHEN 0
  THEN 'CRÍTICO'
  WHEN 1
  THEN 'ALTO'
  WHEN 2
  THEN 'MÉDIO'
  WHEN 3
  THEN 'BAIXO'
END,
CASE (T2318.C1000000163)
  WHEN 1000
  THEN 'EXTENSIVO/DIFUNDIDO'
  WHEN 2000
  THEN 'SIGNIFICATIVO/GRANDE'
  WHEN 3000
  THEN 'MODERADO/LIMITADO'
  WHEN 4000
  THEN 'MENOR/LOCALIZADO'
END,
CASE (T2318.C1000000162)
  WHEN 1000
  THEN 'CRÍTICO'
  WHEN 2000
  THEN 'ALTO'
  WHEN 3000
  THEN 'MÉDIO'
  WHEN 4000
  THEN 'BAIXO'
END
  FROM T2318
  WHERE T2318.C3 > 1434419999;

说明我正在尝试的事情。我将粘贴选择结果的打印

Result of Select

Answer 1

如果你以日期值开头，或者在这种情况下是一个已经转换为日期的值，你可以找到它所属的那一天15分钟的块来操纵午夜过后的秒数;您可以使用to_char()格式模型从SSSSS获取。

select to_char(sysdate, 'YYYY-MM-DD HH24:MI:SS') as now_time,
  to_char(sysdate, 'SSSSS') as now_secs
from dual;

NOW_TIME            NOW_S
------------------- -----
2015-06-18 18:25:49 66349

您可以将秒数向下舍入到15分钟的开始时间除以900（15 * 60），截断或将其放置以获得整数值，然后再乘以900：

select to_char(sysdate, 'YYYY-MM-DD HH24:MI:SS') as now_time,
  to_char(sysdate, 'SSSSS') as now_secs,
  to_number(to_char(sysdate, 'SSSSS'))/900 as calc1,
  floor(to_number(to_char(sysdate, 'SSSSS'))/900) as calc2,
  floor(to_number(to_char(sysdate, 'SSSSS'))/900) * 900 as calc3
from dual;

NOW_TIME            NOW_S      CALC1      CALC2      CALC3
------------------- ----- ---------- ---------- ----------
2015-06-18 18:25:49 66349 73.7211111         73      65700

您可以通过将其添加回日期来将其转换回时间：

select to_char(sysdate, 'YYYY-MM-DD HH24:MI:SS') as now_time,
  to_char(sysdate, 'SSSSS') as now_secs,
  floor(to_number(to_char(sysdate, 'SSSSS'))/900) * 900 as calc3,
  to_char(date '1970-01-01'
    + (floor(to_number(to_char(sysdate, 'SSSSS'))/900) * 900 / 86400),
    'HH24:MI:SS') as calc4
from dual;

NOW_TIME            NOW_S      CALC3 CALC4  
------------------- ----- ---------- --------
2015-06-18 18:25:49 66349      65700 18:15:00

您可能希望保留日期，因此您可以将其添加到trunc(<original_date>)。我想，除非您只在一天内获得数据，或者希望在捆绑在一起的多天显示相同的时间，否则我认为。

这是一个包含10个随机生成时间的演示，显示他们被分配到的15分钟时段：

with t (date_field) as (
  select sysdate - dbms_random.value(0, 1)
  from dual
  connect by level <= 10
)
select to_char(date_field, 'YYYY-MM-DD HH24:MI:SS') as datefield,
  to_char(date_field, 'SSSSS') as time_secs,
  floor(to_number(to_char(date_field, 'SSSSS'))/900) * 900
    as fifteen_min_block_secs,
  to_char(trunc(date_field)
    + (floor(to_number(to_char(date_field, 'SSSSS'))/900) * 900) / 86400,
    'YYYY-MM-DD HH24:MI:SS') as fifteen_min_block
from t
order by datefield;

DATEFIELD           TIME_ FIFTEEN_MIN_BLOCK_SECS FIFTEEN_MIN_BLOCK 
------------------- ----- ---------------------- -------------------
2015-06-17 21:03:00 75780                  75600 2015-06-17 21:00:00
2015-06-18 05:07:28 18448                  18000 2015-06-18 05:00:00
2015-06-18 05:48:42 20922                  20700 2015-06-18 05:45:00
2015-06-18 07:23:03 26583                  26100 2015-06-18 07:15:00
2015-06-18 08:24:57 30297                  29700 2015-06-18 08:15:00
2015-06-18 08:52:06 31926                  31500 2015-06-18 08:45:00
2015-06-18 10:59:14 39554                  38700 2015-06-18 10:45:00
2015-06-18 11:47:05 42425                  42300 2015-06-18 11:45:00
2015-06-18 12:08:37 43717                  43200 2015-06-18 12:00:00
2015-06-18 17:07:23 61643                  61200 2015-06-18 17:00:00

所以你需要拥有

trunc(date_field)
  + (floor(to_number(to_char(date_field, 'SSSSS'))/900) * 900) / 86400

或稍微简单

trunc(date_field)
  + floor(to_number(to_char(date_field, 'SSSSS'))/900) / 96

参与group by子句，可能在您的选择列表中显示。

假设T2318.C3是纪元以来的秒数，您可以直接操作它，然后将其传递给secs_to_datetime函数：

secs_to_datetime(floor(T2318.C3 / 900) * 900)

因此，上述一个等效演示，再次在CTE中随机生成十个时间，将是：

with T2318(c3) as (
  select 1434708000 - dbms_random.value(0, 80000) from dual
  connect by level <= 10
)
select to_char(secs_to_datetime(T2318.C3),'DD/MM/YYYY HH24:MI:SS') as datefield,
  T2318.C3 as time_secs,
  floor(T2318.C3/900) * 900 as fifteen_min_secs,
  to_char(secs_to_datetime(floor(T2318.C3 / 900) * 900),
    'DD/MM/YYYY HH24:MI:SS') as fifteen_min
from T2318
order by T2318.C3;

DATEFIELD              TIME_SECS FIFTEEN_MIN_SECS FIFTEEN_MIN       
------------------- ------------ ---------------- -------------------
18/06/2015 12:34:02   1434630842       1434630600 18/06/2015 12:30:00
18/06/2015 15:06:25   1434639985       1434639600 18/06/2015 15:00:00
18/06/2015 16:43:27   1434645807       1434645000 18/06/2015 16:30:00
18/06/2015 18:57:25   1434653845       1434653100 18/06/2015 18:45:00
18/06/2015 19:01:09   1434654069       1434654000 18/06/2015 19:00:00
18/06/2015 20:54:09   1434660849       1434660300 18/06/2015 20:45:00
19/06/2015 03:59:48   1434686388       1434685500 19/06/2015 03:45:00
19/06/2015 06:58:09   1434697089       1434696300 19/06/2015 06:45:00
19/06/2015 07:36:36   1434699396       1434699000 19/06/2015 07:30:00
19/06/2015 07:47:26   1434700046       1434699900 19/06/2015 07:45:00

或者如果它以毫秒为单位，除以并乘以900000。

Answer 2

在SQL中我们谈论＆＃34;分组＆＃34;我们经常谈论将多行聚合在一起并总结结果。我得到的印象是你只想要＆＃34; round＆＃34;时间值下降到15分钟的开始。数学很简单，但你可能还有一点工作要按照你想要的方式显示它：

FLOOR(EXTRACT(MINUTE FROM datetime_column) / 15) * 15

或者也许：

'00:' || RIGHT('0' || TO_CHAR(FLOOR(EXTRACT(MINUTE FROM datetime_column) / 15) * 15), 2)

我不确定您是否真的有一个日期时间列可以使用，但它与已知格式的字符串值没有太大区别。

FLOOR(TO_NUMBER(RIGHT(HORA_CRIACAO, 2)) / 15) * 15

我想使用整数除法，但我不太确定在Oracle中如何完成。许多人对这种行为感到惊讶，尽管如此，使用FLOOR()的可能性更为明显。

在select

2 个答案: