选择行

Question

我有一个表，看起来像这样：

id person_id element_name element_value element_effective_date 
1 123   a     25    07/01/16 
2 123   b    100    05/01/16 
3 123   a     25    07/20/16 
4 456   b    101    01/01/16 
5 456   b    102    04/01/16 
6 456   a     25    08/01/16 
7 789   b    101    08/01/16 
8 789   c    200    07/25/16 
9 789   d    201    07/26/16 

我想从表中选择的person_ids了selected element_name，其中element_value等于某个值，行的生效日期等于给定日期，或者是person_id的第一行，生效日期小于给定日期。

例如，如果element_name = a且element_value = 25且给定日期为8/5/2016，则所选行将为6和3.如果给定日期更改为7/25/2016，则选定的行会是3.如果日期更改为2016年7月15日，然后选定行是1

Answer 1

​​

Answer 2

数据存储为YYYYMMDD，然后用这个语句：

SELECT * 
FROM table_name 
WHERE element_value = @value 
AND element_name = @name 
AND element_effective_date <= @date 

Answer 3

你的问题并不清楚，无论如何，如果我正确地理解你的问题，下面是解决方案

declare @element_name varchar(1)= 'a' 
declare @element_value int = 25 
declare @element_effective_date smalldatetime ='7/15/2016' 
--'7/25/2016' 
--'8/5/2016' 

create table #a(
id int, 
person_id int, 
element_name varchar(1), 
element_value int, 
element_effective_date smalldatetime) 

insert into #a 
select 1, 123, 'a', 25,'07/01/2016' union 
select 2, 123, 'b', 100,'05/01/2016' union 
select 3, 123, 'a', 25,'07/20/2016' union 
select 4, 456, 'b', 101,'01/01/2016' union 
select 5, 456, 'b', 102,'04/01/2016' union 
select 6, 456, 'a', 25,'08/01/2016' union 
select 7, 789, 'b', 101,'08/01/2016' union 
select 8, 789, 'c', 200,'07/25/2016' union 
select 9, 789, 'd', 201,'07/26/2016' 



select * 
Into #solutionDataset 
from #a 
where element_name=@element_name 
and element_value=@element_value 
and element_effective_date < @element_effective_date 

/* 
I want to select the person_ids from the table for a selected element_name where the element_value is equal to some value and the rows' effective date is equal to a given date or is the first row for the person_id with the effective date less than the given date. 
*/ 
if(@@ROWCOUNT=1) 
begin 
select * from #solutionDataset 
end 

else 
begin 
    select * from #solutionDataset 
    where id not in(select min(id) from #a where element_name=@element_name and element_value=@element_value) 
end 

Answer 4

这工作：

select person_id 
from 
(
    select max(element_effective_date), person_id 
    from 
    (
     select t.person_id, t.element_effective_date 
     from table t 
     where t.element_name = 'a' and t.element_value = 25 
      and t.element_effective_date <= '2016-08-05' 
    ) as foo 
    group by person_id 
); 

THX您的意见。

Answer 5

不确定我是否遵循了所有你想要的逻辑，但是这些应该是非常接近的。我也假设SQL Server，但这只应该参数名称。

如果你能保证id值都在增加：

select * 
from <T> 
where id in (
    select max(id) 
    from <T> 
    where element_value = @element_value and element_effective_date <= @effective_date 
    group by person_id 
) 

如果没有，但你仍然可以保证effective_date无关系：

select t.* 
from 
    <T> t 
    inner join 
    (
     select person_id, element_value, max(element_effective_date) as max_date 
     from <T> 
     where element_effective_date <= @effective_date 
     group by person_id 
    ) as m 
     on  m.person_id = t.person_id 
      and m.element_value = t.element_value 
      and m.max_date = t.element_effective_date 
where 
    t.element_value = @element_value