使用group by子句

时间:2020-06-25 10:33:48

标签: mysql sql database group-by

我正在使用Mysql数据库,并且有一个时隙表,其中应将基于允许计数(N)记录的特定日期的不同时隙更新为有效。

CREATE  TABLE tmpSlots(
    SlotID INT AUTO_INCREMENT PRIMARY KEY  ,
    StartSlot DATETIME,
    EndSlot DATETIME,
    Valid BOOLEAN DEFAULT 0
);  
insert into tmpSlots VALUES(1,'2020-08-01 08:30:00', '2020-08-01 09:10:00',0 );
insert into tmpSlots VALUES(2,'2020-08-01 09:30:00', '2020-08-01 10:10:00',0 );
insert into tmpSlots values(3,'2020-08-01 10:30:00', '2020-08-01 11:10:00',0 );
insert into tmpSlots values(4,'2020-08-01 12:30:00', '2020-08-01 13:10:00',0 );
insert into tmpSlots values(5,'2020-08-07 08:30:00', '2020-08-07 09:10:00',0 );
insert into tmpSlots values(6,'2020-08-07 09:30:00', '2020-08-07 10:10:00',0 );
insert into tmpSlots values(7,'2020-08-07 10:30:00', '2020-08-07 11:10:00',0 );
insert into tmpSlots values(8,'2020-08-07 12:30:00', '2020-08-07 13:10:00',0 );

DECLARE permitcount INT ; 
SET permitcount =2;

由于allowcountcount,因此每天只允许2个广告位,因此前2条记录应更新为有效= true 预期结果 enter image description here

UPDATE tmpSlots t1 SET valid=1
FROM  (
  ...........
.
   GROUP  BY Date(StartSlot)
   ) AS sq
WHERE 

有人可以帮我吗

1 个答案:

答案 0 :(得分:1)

您可以在join中使用update。但是,您不希望聚合。相反,您可以使用row_number()来枚举每天的值。然后,使用where子句选择每天的前两个:

update tmpslots s join
       (select s2.*,
               row_number() over (partition by date(startslot) order by startslot) as seqnum
        from tmpslots s2
       ) s2
       on s2.slotid = s.slotid
    set s.value = 1
    where s2.seqnum <= 2;

这根据startslot分配日期。

在早期版本中,我将仅使用两个更新:

update tmpslots s join
       (select date(startslot) as dte, min(slotid) as min_slotid
        from tmpslots s2
        group by dte
       ) s2
       on s2.min_slotid = s.slotid
    set s.value = 1;

update tmpslots s join
       (select date(startslot) as dte, min(slotid) as min_slotid
        from tmpslots s2
        where s.value = 0
        group by dte
       ) s2
       on s2.min_slotid = s.slotid
    set s.value = 1;

尽管您可以将其添加到单个更新中,但两个更新似乎更简单。

相关问题