通过 Where in 获取每个 Id 的前 1 行答案

【问题标题】：Get top 1 row of each Id by Where in通过 Where in 获取每个 Id 的前 1 行
【发布时间】：2020-01-13 02:54:13
【问题描述】：

我有一个表，我想在其中获取每个参数的最新条目

select GpsData.VehicleId, GpsData.Id, GpsData.DateTime
from GpsData
where GpsData.VehicleId in (44, 1054, 1055, 31, 22, 1058)
order by GpsData.VehicleId desc;

【问题讨论】：

标签： sql-server tsql query-performance where-in

【解决方案1】：

与NOT EXISTS:

select g.VehicleId, g.Id, g.DateTime
from GpsData g
where g.VehicleId in (44, 1054, 1055, 31, 22, 1058)
and not exists (
  select 1 from GpsData
  where VehicleId = g.VehicleId and DateTime > g.DateTime 
)
order by g.VehicleId desc;

或者用row_number()窗口函数：

select t.VehicleId, t.Id, t.DateTime
from (
  select VehicleId, Id, DateTime,
    row_number() over (partition by VehicleId order by DateTime desc) as rn   
  from GpsData 
  where VehicleId in (44, 1054, 1055, 31, 22, 1058)
) as t
where t.rn = 1
order by t.VehicleId desc;

【讨论】：

我更喜欢第二种方法...我会将FROM(...) 转换为CTE，但这只是语法糖。从我这边 +1。

【解决方案2】：

您可以使用TOP 1 WITH TIES

SELECT TOP 1 WITH TIES
    VehicleId, 
    Id, 
    [DateTime]
FROM GpsData
WHERE VehicleId in (44, 1054, 1055, 31, 22, 1058)
ORDER BY ROW_NUMBER() OVER (PARTITION BY VehicleId ORDER BY [DateTime] DESC)

我们可以使用动态查询如下：

 DECLARE @Ids NVARCHAR(MAX) = '44, 1054, 1055, 31, 22, 1058', @sql NVARCHAR(3000)

 SET @sql = '
    SELECT TOP 1 WITH TIES VehicleId, 
        Id, 
        [DateTime] 
    FROM GpsData 
    WHERE VehicleId IN (' + @ids + ') 
    ORDER BY ROW_NUMBER() OVER (PARTITION BY VehicleId ORDER BY [DateTime] DESC)'

EXEC(@sql)

【讨论】：

Susang，我真的很喜欢这种巧妙的方法，但我不得不发现，它非常慢...排序将执行两次（一次用于ROW_NUMBER，一次用于ORDER BY . 较大的集合相当不利...最好使用CTE和ROW_NUMBER作为列并在最后使用WHERE RNr=1...
非常感谢。我如何为这种方法使用参数声明@Ids nvarchar(Max) =44, 1054, 1055, 31, 22, 1058 SELECT TOP 1 WITH TIES VehicleId, Id, [DateTime] FROM GpsData WHERE VehicleId in (@ids) ORDER BY ROW_NUMBER() OVER（PARTITION BY VehicleId ORDER BY [DateTime] DESC）
我们可以使用上面更新的动态查询，请检查。