SELECT COUNT指数

Question

我是一个MySQL数据库优化的新手。

我有以下慢速查询：

SELECT COUNT(DISTINCT p.product_id) AS total 
    FROM shs_product p 
LEFT JOIN shs_product_description pd 
     ON (p.product_id = pd.product_id) 
LEFT JOIN shs_product_to_store p2s 
     ON (p.product_id = p2s.product_id) 
LEFT JOIN shs_product_to_category p2c 
     ON (p.product_id = p2c.product_id) 
    WHERE pd.language_id = '1' 
     AND p.status = '1' 
     AND p.date_available <= NOW() 
     AND p2s.store_id = '$ 

谁能行动提供建议创建索引，以加快此查询？ 您推荐哪张桌子和哪一列？

任何帮助将不胜感激......

Answer 1

三点意见：

1. 的where子句撤消外连接p2s和pd。所以你不妨称这些内部连接。
2. p2c表未被使用。因为您正在执行count(distinct)，您可以将其删除。
3. 据推测，该ID是整数，所以你不需要引号

因此，查询是等效

SELECT COUNT(DISTINCT p.product_id) AS total 
FROM shs_product p JOIN 
    shs_product_description pd 
    ON p.product_id = pd.product_id JOIN 
    shs_product_to_store p2s 
    ON p.product_id = p2s.product_id 
WHERE pd.language_id = 1 AND p.status = 1 AND p.date_available <= NOW() AND 
     p2s.store_id = '$' 

对于此查询，您想对shs_product(status, date_available)的索引。您还需要用于加入的列和where子句中的索引。我会建议：shs_product_description(product_id, language_id)和shs_product_to_store(product_id, store_id)。

最后，假设product_id是一个独特的指数product，您可以短语使用exists和count(*)，而不是count(distinct)此查询：

select count(*) 
from shs_product p 
where p.status = 1 AND p.date_available <= NOW() and 
     exists (select 1 
       from shs_product_description pd 
       where p.product_id = pd.product_id and pd.language_id = 1 
      ) and 
     exists (select 1 
       from shs_product_to_store p2s 
       where p.product_id = p2s.product_id and p2s.store_id = '$' 
      );