Question

我有一个union查询，我希望在union查询的“右侧”的select语句中使用union查询的“左侧”中的select结果。下面的查询正常工作（至少在postgres上），但我运行query1 2次，一次作为query1，再次作为sameAsQuery1。

select x as zz from (select 69 as x) as query1
union all
select count(zz) as zz from 
      (select x as zz from (select 69 as x) as sameAsQuery1) as query2

我想做这样的事情，所以我不必运行query1 2次，但它不起作用：

select x as zz from (select 69 as x) as query1
union all
select count(zz) as zz from query1

我收到此错误消息：

错误：关系“query1”不存在第3行：选择计数（zz）作为来自query1的zz

有没有办法重写这个查询，所以query1只运行一次？

对Llama先生的回应进行了一些修改，效果很好，看起来像这样（注意添加“为q2”）：

WITH
query1 AS
(
    SELECT x AS zz FROM (SELECT 69 AS x) as q2
)
SELECT zz FROM query1
UNION ALL
SELECT COUNT(zz) AS zz FROM query1

Answer 1

您正在寻找common table expressions。

它们允许您定义结果并在查询中多次使用它。

在你的第一种情况下：

WITH
query1 AS
(
    SELECT x AS zz FROM (SELECT 69 AS x)
)
SELECT zz FROM query1
UNION ALL
SELECT COUNT(zz) AS zz FROM query1

Answer 2

当大多数数据库支持group by和rollup时，为什么会这样做？看起来你想要这样的东西：

select x, count(*) as cnt
from <whatever>
group by x with rollup;

Answer 3

将query1结果放入临时表

select x as zz from (select 69 as x) as query1 into temptable

现在在第二个查询中使用临时表

select zz from temptable
union all
select count(zz) as zz from temptable

Answer 4

您可能需要使用公用表表达式来重用SELECT：

with cte as
( select x as zz from (select 69 as x) as query1 )
select * from cte
union all
select count(zz) as zz from cte

如何在联合查询中的另一个select中使用select的结果？

4 个答案: