PostgreSQL 实现子查询返回多行的案例

Z时代
2024-01-10
分类：综合

作为一个表达式使用的子查询返回了多列：

在查询中，我们需要以第2条查询语句作为第一条查询语句的条件，但是第一条根据这个条件查询出来是多个数据，这时候又需要保留多个数据，运用子查询就会报错，

以下就为解决这种多对多关系查询，且没有关联关系的表的解决方案：


select c.rain_value,c.ad_nm from 
(
select *, json::json->t2.lon_lat as rain_value from actual_time_model_json t1,
(SELECT DISTINCT lon || '_' || lat as lon_lat,ad_nm from grid_all_points_null)t2 where section='0' and t1.filename = 'Z_NWGD_C_BCCD_20180711022009_P_RFFC_SPCC-ER01_201807110800_02401.GRB2'
)c where c.rain_value is not null

补充：PostgreSQL 的子查询相关的知识，exists,any,all

Subquery


SELECT
 film_id,
 title,
 rental_rate
FROM
 film
WHERE
 rental_rate > (
  SELECT
   AVG (rental_rate)
  FROM
   film
 );

有了子查询，在设定需要查询表才能得到查询条件时，就可以直接在一条语句中写，不用分开多条写了，方便了许多。

子查询返回多条时，可以在 where 子句中用 IN，来匹配查询条件。

SELECT

film_id,

title

FROM

film

WHERE

film_id IN (

SELECT

inventory.film_id

FROM

rental

INNER JOIN inventory ON inventory.inventory_id = rental.inventory_id

WHERE

return_date BETWEEN '2005-05-29'

AND '2005-05-30'

);

EXISTS 操作符

在 where 子句的查询条件中，exists 操作符，会在子查询有返回行时，返回true；不论返回几行。

因此，子查询中的查询字段仅写1就好；标准的写法：EXISTS (SELECT 1 FROM tbl WHERE condition)


SELECT
 first_name,
 last_name
FROM
 customer
WHERE
 EXISTS (
  SELECT
   1
  FROM
   payment
  WHERE
   payment.customer_id = customer.customer_id
 );

NO EXISTS ,与之相反，当子查询返回0行时，返回true


SELECT first_name,
  last_name
FROM customer c
WHERE NOT EXISTS
 (SELECT 1
  FROM payment p
  WHERE p.customer_id = c.customer_id
  AND amount > 11 )
ORDER BY first_name,
   last_name;

当子查询返回 NULL，会返回true, 也就是返回所有行。


SELECT
 first_name,
 last_name
FROM
 customer
WHERE
 EXISTS( SELECT NULL )
ORDER BY
 first_name,
 last_name;

ANY

与任何子查询返回的值匹配就返回 true

expresion operator ANY(subquery)

表达式一般为字段

操作符为 >,<,=,<>,>=,<=

ANY 可以与 SOME 替换

子查询返回的必须是一列，


SELECT title
FROM film
WHERE length >= ANY(
 SELECT MAX( length )
 FROM film
 INNER JOIN film_category USING(film_id)
 GROUP BY category_id );

The = ANY is equivalent to IN operator.

Note that the <> ANY operator is different from NOT IN. The following expression:


x <> ANY (a,b,c)

is equivalent to


x <> a OR x <> b OR x <> c

ALL

所有子查询返回的值匹配则返回 true

也就是大于最大，小于最小


SELECT
 film_id,
 title,
 length
FROM
 film
WHERE
 length > ALL (
   SELECT
    ROUND(AVG (length),2)
   FROM
    film
   GROUP BY
    rating
 )
ORDER BY
 length;