按列分组并获取另一列中具有最大字符串长度的行

Question

我在 Postgres 11 中关注 table。

table1:

id          col1    col2                        col3                        col4                        
NCT00000374 Drug    olanzapine                  olanzapine                  olanzapine                  
NCT00000390 Drug    imipramine hydrochloride    imipramine hydrochloride    imipramine hydrochloride    
NCT00000390 Drug    imipramine hydrochloride    imipramine hydrochloride    imipramine                  
NCT00000412 Drug    placebo calcitriol          placebo calcitriol          calcitriol

我想获取每个 (id, col1, col2, col3).

具有最大长度值的行

期望的输出是：

id          col1    col2                        col3                        col4                        
NCT00000374 Drug    olanzapine                  olanzapine                  olanzapine                  
NCT00000390 Drug    imipramine hydrochloride    imipramine hydrochloride    imipramine hydrochloride    
NCT00000412 Drug    placebo calcitriol          placebo calcitriol          calcitriol

到目前为止，我尝试了以下查询但没有成功：

select * from table1
where length(col4) = max(length(col4))
group by id, col1, col2, col3
order by id

Answer 1

DISTINCT ON的案例：

SELECT DISTINCT ON (id, col1, col2, col3)
       *
FROM   table1
ORDER  BY id, col1, col2, col3, length(col4) DESC NULLS LAST;

最简单，每行几行 (id, col1, col2, col3) 通常也是最快的。详细解释：

Select first row in each GROUP BY group?

对于大表和每组很多行，有（多）更快的技术：

Optimize GROUP BY query to retrieve latest row per user

按列分组并获取另一列中具有最大字符串长度的行

Group by columns and fetch rows with maximum length of string in another column

sql

postgresql

aggregation

greatest-n-per-group