我正在尝试从数据集中选择所有记录,其中1条记录符合条件。我有一张数据表,其中包含一个人每天的工作时间。如果某个人在该周内的一条记录的"logged"标志设置为"N",则我需要选择该周的所有记录。下面是一个数据示例:
t_hours:
Name Week Weekday Hours Logged
===============================
Jim 1 Mon 8 Y
Jim 1 Wed 8 Y
Jim 1 Fri 8 Y
Jim 2 Mon 8 Y
Jim 2 Wed 8 Y
Bill 1 Mon 8 N
Bill 1 Tue 8 Y
Bill 1 Wed 8 Y
Bill 1 Thu 8 Y
Bill 2 Mon 8 Y
Bill 2 Tue 8 Y
我想写一个查询,如果有一天Logged='N',它会将一个人工作周的所有记录更新为Logged='N+。但我甚至不知道如何选择记录。以下是我想更新的记录:
Name Week Weekday Hours Logged
===============================
Bill 1 Mon 8 N
Bill 1 Tue 8 Y
Bill 1 Wed 8 Y
Bill 1 Thu 8 Y
我尝试了一个普通的选择,但不知道如何在where子句中有两个相关的子查询:
SELECT * FROM t_hours
WHERE (Name = (SELECT t1.Name FROM t_hours t1
where t1.Name = t2.Name and t1.Week = t2.Week and
t1.Logged = 'N') and
Week = (SELECT t2.Week FROM t_hours t2
where t1.Name = t2.Name and t1.Week = t2.Week and
t2.Logged = 'N')
但这不起作用,我们非常感谢您的帮助。
您可以使用可更新的CTE:
with toupdate as (
select t.*,
min(logged) over (partition by name, week) as min_logged
from t
)
update toupdate
set logged = min_logged
where min_logged = 'N' and min_logged <> logged;
如果某个人/周的logged
值中的任何一个为'N'
(如果有的话,其余值为'Y'
(,则min(logged)
表达式将返回'N'
。
返回数据:
SELECT *
FROM t_hours t1
WHERE EXISTS (SELECT 1
FROM t_hours t2
WHERE t1.NAME = t2.NAME
AND t1.week = t2.week
AND t2.logged = 'N');
更新数据:
UPDATE t_hours
SET logged = 'N'
WHERE EXISTS (SELECT 1
FROM t_hours t2
WHERE t1.NAME = t2.NAME
AND t1.week = t2.week
AND t2.logged = 'N');
如果数据集足够大,可以按名称、周和日志类型进行分组。
使用现有进行如下尝试
select t1.* from table_name t1
where exists( select 1 from table_name t2 where t1.week=t2.week
and t2.logged='N' and t1.name=t2.name
)
这就是您想要的吗?
Update table set Logged='N' where
Exists (Select 1 from(Select
Name,Workweek group by Name,
Workweek having count(Logged='N')
>=1)t1 where Name=t1.name and
Workweek=t1.Workweek)