MySQL存在使查询耗时太长

Question

我对MySQL很新，现在我尝试进行以下选择：

SELECT
        a.qtd,
        a.value
    FROM
        order_products a,
        products b,
        orders c
    WHERE
        b.id = a.product_id
        and a.order_id = c.id
        and c.status = '2'
        and c.seller_id = '$result->seller_id'
        and exists(select id from dealer_products p where p.product_id = b.id and p.id_dealer = '$dealer')

此查询使用LIMIT '1'大约需要58秒，如果没有，我甚至不会得到回复。

删除and exists(select id from dealer_products p where p.product_id = b.id and p.id_dealer = '$dealer')会使响应时间降至0.78秒，以返回相同的1个结果。

这是我在select上使用EXPLAIN时得到的结果：

有没有办法改进这个查询，以便它不会花费这么荒谬的时间？

Answer 1

exists(select id from dealer_products p where p.product_id = b.id and p.id_dealer = '$dealer')

而不是你可以使用

b.id in (select id from dealer_products p where p.product_id = b.id and p.id_dealer = '$dealer')

Answer 2

亲爱的回答指出了我正确的方向。

仍然使用select([t]).where(getattr(t.c, col_name1) == "foo") 未返回正确的值。

要解决此问题，我已将其更改为b.id in (select id from dealer_products p where p.product_id = b.id and p.id_dealer = '$dealer')。工作得很好，响应时间0.012秒。

Answer 3

请使用JOIN ... ON语法：

SELECT  ...
    FROM  orders         AS o
    JOIN  order_products AS op  ON op.order_id = o.id
    JOIN  products       AS  p  ON p.id = op.product_id
    WHERE  o.status = 2
      AND  o.seller_id = '...'
      AND  EXISTS(
               SELECT  *    -- Note: no need for `id`
                   FROM  dealer_products
                   WHERE  product_id = p.id
                     AND  id_dealer = '...' 
                 )

所需索引：

orders:  INDEX(status, seller_id, id) -- in either order
order_products:  INDEX(order_id, product_id)  -- "covering"
products:  INDEX(id)  -- but not if it is already PRIMARY KEY(id)
dealer_products: INDEX(id_dealer, product_id) -- note slight change in subquery

另外，请按照Many:Many mapping tables中的提示进行操作。