Question

我正在尝试优化用于电子商务-自定义商店的mysql查询，该查询具有一些内部联接和分组依据，计数使查询非常慢，我们使用它来生成类别的动态过滤器/组合，供应商，颜色等。

有人对如何优化它有任何想法吗？首先，我无法使products（p）表使用多重索引而不是主键，查询大约需要150-200ms。

确切的查询：

SELECT DISTINCT(sl.id_supplier), 
       count(DISTINCT p.id_product) as product_count, 
       sl.supplier_name, 
       sl.supplier_slug, 
       sl.supplier_link 
FROM supplier s 
INNER JOIN supplier_lang sl 
  on s.id_supplier = sl.id_supplier 
  AND sl.id_project = 6 
  AND sl.id_lang =2 
  AND s.id_project = 6 
  AND s.active = 1 
INNER JOIN product p 
  ON p.id_project = 6 
  and p.active = 1 
  AND p.id_supplier = sl.id_supplier 
INNER JOIN product_category pc 
  on pc.id_project = 6 
  and pc.id_category = 6 
  and p.id_product = pc.id_product  
GROUP by sl.id_supplier 
ORDER BY sl.supplier_name ASC

谢谢

https://snag.gy/F6dy4r.jpg

Answer 1

尽管上面有隐式的函数依赖关系，但我认为这样的查询看起来更像是一种好习惯（通常会期望）：

SELECT sl.id_supplier
     , sl.supplier_name
     , sl.supplier_slug
     , sl.supplier_link 
     , COUNT(DISTINCT p.id_product) product_count 
  FROM supplier s 

  JOIN supplier_lang sl 
    ON sl.id_supplier = s.id_supplier 
   AND sl.id_project = s.id_project 
   AND sl.id_lang = 2 

  JOIN product p  
    ON p.id_supplier = sl.id_supplier 
   AND p.id_project = s.id_project  
   AND p.active = 1 

  JOIN product_category pc 
    ON pc.id_product = p.id_product 
   AND pc.id_project = p.id_project 
   AND pc.id_category = 6 

 WHERE s.id_project = 6 
   AND s.active = 1 

 GROUP 
    BY sl.id_supplier 
     , sl.supplier_name
     , sl.supplier_slug
     , sl.supplier_link 
 ORDER 
    BY sl.supplier_name ASC

为帮助优化此查询，我们将需要查看所有相关表的SHOW CREATE TABLE语句以及上述内容的EXPLAIN。这将是您的问题的编辑形式，并且将不包含任何图片。

MySQL多组和顺序很慢

1 个答案: