带有 case 语句的 having 子句

having clause with case statement

我正在尝试获取具有最大日期 col4_date 的行分组,其中 col5 = Hi。我注意到其他人对案例进行了类似的查询。也许我需要使用 in 子句来执行此操作。最大值 (col4_date) 是失败的地方。任何建议都会很好

select  col1, col2, col3  from table1 t1 
group by col1, col2, col3
HAVING 
 sum(case when ( max (col4_date) =  col4_date ) and (col5 = 'Hi' )  then 1 else 0 end) > 0

table 1 个内容

col1 | col2 | col3 | col4_date | col5
-----+------+------+-----------+-----
D    | F    | G    | 4/3/2018  | Hi 
D    | F    | G    | 1/1/1970  | Bye
H    | I    | J    | 1/1/1970  | Hi 
H    | I    | J    | 4/3/2018  | Bye

输出

col1 | col2 | col3
-----+------+-----
D    | F    | G   

感谢您的帮助

这可能就是你想要的

select  col1, col2, col3  
from table1 t1
join (select col1, col2, col3, max(col4_date) as max_col4_date
      from table1
      group by col1, col2, col3
) sub on sub.col1 = t1.col1 and sub.col2 = t1.col2 and sub.col3 = t1.col3 and sub.max_col4_date = t1.col4_date
where col5 = 'Hi'

写作sub-query作品

SELECT col1, col2, col3
  FROM grp
 WHERE     (col1, col2, col3, col4) IN (  SELECT col1,
                                                 col2,
                                                 col3,
                                                 MAX (col4)
                                            FROM grp
                                        GROUP BY col1, col2, col3)
       AND col5 = 'Hi';

使用 Oracle 的 KEEP LAST 获取最后 col4_datecol5 值:

select col1, col2, col3
from table1
group by col1, col2, col3
having max(col5) keep (dense_rank last order by col4_date) = 'Hi';

这应用了一个 CASE 来得到你的结果:

select  col1, col2, col3
from table1 t1 
group by col1, col2, col3
HAVING -- compare overall max date and max date for 'Hi'
   max(col4_date) = max(case when col5 = 'Hi' then col4_date end)