Sql 获取有效的日期范围。最小值和最大值

Sql 获取有效的日期范围。最小值和最大值,sql,postgresql,validation,pentaho,pentaho-data-integration,Sql,Postgresql,Validation,Pentaho,Pentaho Data Integration,我需要根据有效性范围获取表的值。目前,我根据表格的列进行分组,并对该年的最小值和最大值进行分组;但对于这个例子,这给了我错误的结果。我需要用sql语句或pentaho来解决它 样本数据: year validity Value1 Value2 2004 A B 2006 A C 2007 A B 2008 A B SQL: 错误结果: year

我需要根据有效性范围获取表的值。目前,我根据表格的列进行分组,并对该年的最小值和最大值进行分组;但对于这个例子,这给了我错误的结果。我需要用sql语句或pentaho来解决它

样本数据:

year validity   Value1  Value2
2004               A     B
2006               A     C
2007               A     B
2008               A     B
SQL:

错误结果:

year min    year max    Value1  Value2
2004        2008           A      B
2006        2006           A      C
预期结果:

year min    year max    Value1  Value2
2004        2004         A        B
2007        2008         A        B
2006        2006         A        C

请在此问题上提供帮助。

您可以组合
LAG()
SUM()
窗口函数,根据您的逻辑将行分成若干组。一旦分离,计算最小值和最大值就很简单了

例如:

select
  min(year_validity) as year_min,
  max(year_validity) as year_max,
  min(value1) as value1,
  min(value2) as value2
from (
  select
    *,
    sum(init) over(order by year_validity) as grp
  from (
    select *,
    case when not ( (value1, value2) = (
      lag(value1) over(order by year_validity),
      lag(value2) over(order by year_validity)
    ) ) then 1 else 0 end as init
    from tabla
  ) x
) y
group by grp
order by value1, value2
结果:

year_min  year_max  value1  value2
--------  --------  ------  ------
2004      2004      A       B
2007      2008      A       B
2006      2006      A       C
对于记录,本案例的数据脚本为:

create table tabla (
  year_validity int,
  value1 varchar(10),
  value2 varchar(10)
);

insert into tabla (year_validity, value1, value2) values (2004, 'A', 'B');
insert into tabla (year_validity, value1, value2) values (2006, 'A', 'C');
insert into tabla (year_validity, value1, value2) values (2007, 'A', 'B');
insert into tabla (year_validity, value1, value2) values (2008, 'A', 'B');
另一个“缺口和孤岛”问题。
create table tabla (
  year_validity int,
  value1 varchar(10),
  value2 varchar(10)
);

insert into tabla (year_validity, value1, value2) values (2004, 'A', 'B');
insert into tabla (year_validity, value1, value2) values (2006, 'A', 'C');
insert into tabla (year_validity, value1, value2) values (2007, 'A', 'B');
insert into tabla (year_validity, value1, value2) values (2008, 'A', 'B');