SQL 可以简化吗？

Question

有人问我以下关于 table 的问题

Display NPRO and LIBELLE from SAPIN products that are ordered only in Toulouse

这是我的答案（正确），但我觉得可以以某种方式简化它。

有没有办法让它更简单？

select npro, libelle
from produit
where libelle like '%SAPIN%'
and npro in 
(select npro from detail where ncom in
(select ncom from commande where ncli in
(select ncli from client where localite='Toulouse')))
and npro not in
(select npro from detail where ncom in
(select ncom from commande where ncli in
(select ncli from client where localite<>'Toulouse')))

Answer 1

一种使其更易于阅读的方法是使用：

SELECT p.npro, p.libelle
FROM produit p
JOIN detail d
  ON p.npro = d.npro
JOIN commande c
  ON d.ncom = c.ncom
JOIN client cl
  ON cl.ncli = c.ncli
WHERE p.libelle LIKE '%SAPIN%'
GROUP BY p.npro, p.libelle
HAVING COUNT(CASE WHEN cl.localite = 'Toulouse' THEN 1 END) > 0
   AND COUNT(CASE WHEN cl.localite <> 'Toulouse' THEN 1 END) = 0

工作原理：

加入您需要的表
从包含 SAPIN

p.libelle

按 p.npro 和 p.libelle
聚合过滤结果：第一个计数 localite 等于 Touluse，第二个计数 localite.

Answer 2

EXISTS+NOT EXISTS 最简单。（恕我直言）

SELECT npro, libelle
FROM produit p
WHERE p.libelle LIKE '%SAPIN%'
AND EXISTS (
        SELECT *  
        FROM detail d  
        JOIN commande co ON co.ncom = d.ncom
        JOIN client cl ON cl.ncli = co.ncli
        WHERE d.npro = p.npro
        AND cl.localite = 'Toulouse'
        )
AND NOT EXISTS (
        SELECT *
        FROM detail d
        JOIN commande co ON co.ncom = d.ncom
        JOIN client cl ON cl.ncli = co.ncli
        WHERE d.npro = p.npro
        AND cl.localite <> 'Toulouse'
        );

SQL 可以简化吗？

SQL simplification possible?

sql

postgresql

relational-division