Question

出于性能目的，是否可以在SELECT列表中重用WHERE子句中的子选择，如：

SELECT p.id, nItens FROM purchase AS p 
WHERE ((SELECT COUNT(*) AS n FROM itens WHERE purchase_id=p.id) AS nItens) > 1

避免：

SELECT p.id, (SELECT COUNT(*) AS n FROM itens WHERE purchase_id=p.id) AS nItens 
FROM purchase AS p 
WHERE (SELECT COUNT(*) AS n FROM itens WHERE purchase_id=p.id) > 1

Answer 1

SELECT p.id, (SELECT COUNT(*) AS n FROM itens WHERE purchase_id=p.id) AS nItens 
FROM purchase AS p 
HAVING nItens  > 1

Answer 2

SELECT p.id
     , COUNT(n.purchase_id) ttl
  FROM purchase p 
  JOIN itens n 
    ON n.purchase_id = p.id
 GROUP
    BY p.id
HAVING ttl > 1

Answer 3

在MySQL中，您可以在having子句中使用列别名，但不能在where子句中使用列别名。如果你在MySQL中有一个没有having的{{1}}子句 - 那么它的行为就像一个group by。所以你可以这样做：

where

但是，您的查询可能更好地写为：

SELECT p.id, (SELECT COUNT(*) AS n FROM itens WHERE purchase_id=p.id) AS nItens 
FROM purchase AS p 
HAVING nItens  > 1;

请注意，如果没有条件，您需要一个select p.id, count(*) as nItens from purchase p join itens i on p.id = i.purchase_id group by p.id having count(*) > 1;来获得相同的逻辑，以保持与之无匹配的购买。由于您的条件至少需要两次此类匹配，因此您不需要左连接。

如何在SELECT和WHERE子句中使用相同的子选择

3 个答案: