Question

Select Distinct top 100  Uid,ProDate,ProTime,min(Auto_No) As Auto_No
from Has_Swipedata 
where  format(Prodate, 'dd/MM/yyyy') = Format(Getdate(), 'dd/MM/yyyy') and
       Is_Active='N' and
       Uid not like '%[^0-9]%' and Uid != '' 
Group by Uid, ProDate, ProTime

来自上方查询的表数据： Uid ProDate ProTime Auto_No

4057    2017-06-14  16:51:12.0000000    1584

4057    2017-06-14  16:51:13.0000000    1585

4057    2017-06-14  16:51:15.0000000    1586

预期结果：

Uid      ProDate    ProTime             Auto_No

4057    2017-06-14  16:51:12.0000000    1584

Answer 1

您无需将日期格式化为字符串即可比较日期。如果ProDate已经是date数据类型，请与convert(date,getdate())进行比较。

这会得到每个Auto_No和Uid的第一个/分钟ProDate，以及ProTime的{{1}}。

使用Auto_No：

cross apply()

或使用内部联接：

select distinct top 100 
    sd.Uid
  , sd.ProDate
  , x.ProTime
  , x.Auto_No as Auto_No
from Has_Swipedata sd
  cross apply (
    select top 1 *
    from Has_Swipedata i
    where i.Uid = sd.UID
      and i.ProDate = sd.ProDate
    order by i.Auto_No asc
    ) x
where sd.ProDate = convert(date,getdate())
  and sd.Is_Active = 'N'
  and sd.Uid not like '%[^0-9]%'
  and sd.Uid != ''

Answer 2

看看想要的结果，我猜是

Select top(1) with ties Uid, ProDate, ProTime, Auto_No 
from Has_Swipedata 
where  format(Prodate, 'dd/MM/yyyy') = Format(Getdate(), 'dd/MM/yyyy') and
       Is_Active='N' and
       Uid not like '%[^0-9]%' and Uid != '' 
order by row_number() over(partition by Uid,ProDate order by Auto_No)

演示

select top(1) with ties Uid, ProDate, ProTime, Auto_No 
from 
(values
 (4057,    '2017-06-14',  '16:51:12.0000000',   1584),
 (4057,    '2017-06-14',  '16:51:13.0000000',   1585),
 (4057,    '2017-06-14',  '16:51:15.0000000',   1586),
 (4058,    '2017-06-14',  '16:51:12.0000000',   1587),
 (4058,    '2017-06-14',  '16:51:13.0000000',   1588),
 (4058,    '2017-06-14',  '16:51:15.0000000',   1589)
) Has_Swipedata (Uid,ProDate,ProTime, Auto_No)
order by row_number() over(partition by Uid,ProDate order by Auto_No)

返回

    Uid ProDate ProTime Auto_No
1   4057    2017-06-14  16:51:12.0000000    1584
2   4058    2017-06-14  16:51:12.0000000    1587

如果您只需要前100个Uid + ProDate，请使用Select top（100）包装查询......

select top(100) *
from (
    select top(1) with ties Uid, ProDate, ProTime, Auto_No 
    from 
    (values
     (4057,    '2017-06-14',  '16:51:12.0000000',   1584),
     (4057,    '2017-06-14',  '16:51:13.0000000',   1585),
     (4057,    '2017-06-14',  '16:51:15.0000000',   1586),
     (4058,    '2017-06-14',  '16:51:12.0000000',   1587),
     (4058,    '2017-06-14',  '16:51:13.0000000',   1588),
     (4058,    '2017-06-14',  '16:51:15.0000000',   1589)
    ) Has_Swipedata (Uid,ProDate,ProTime, Auto_No)
    order by row_number() over(partition by Uid,ProDate order by Auto_No)
) result

Answer 3

也许在ProTime中应用相同的Auto_No规则会生成预期结果：

Select top 100  
    Uid, ProDate, min(ProTime) as ProTime, min(Auto_No) As Auto_No
from Has_Swipedata 
where  Prodate = convert(date, GETDATE()) and
       Is_Active='N' and
       Uid not like '%[^0-9]%' and Uid != '' 
Group by Uid, ProDate

或提供更多具有预期结果的数据样本。

明确不适用于以下查询

3 个答案: