如何将count（*）引用为where子句中的变量以及group by子句

Question

我正在使用DB2并尝试限制使用group by的查询仅包含大于某个阈值的计数但无法通过语法错误

select val, count(*) as c from sample_table
where c > 20
group by val;

它不喜欢我引用c并在where子句中使用它的事实。

我如何实现这样的目标？

Answer 1

你可以这样做：

select *
from (
select val, count(*) as c 
from sample_table
group by val
)
where c > 20

但要做得更好：

select val, count(*) as c 
from sample_table
group by val
having count(*) > 20

Answer 2

您可以使用HAVING：

select val, count(*) as c 
from sample_table
group by val;
having count(*)> 20

Answer 3

你不能在where子句中使用聚合函数，所以你必须使用HAVING。

select val, count(*) as c
from sample_table
group by val
Having c>20;