MySQL的优化

Question

我有mysql-表称为产品具有以下结构/内容：

id (int), public_id (int), deleted (bool), online (bool) 
1   1    0    1 
2   1    0    0 
3   1    1    0 
4   2    0    1 
5   2    0    1 

我的问题是，如何选择全部，当前在线，不删除产品。在这个例子中只记录5（public_id 2）。相同的public_id表示相同的产品（分组）。id越高，信息越新（排序）。并且产品需要而不是已删除（其中）。还有一些其他地方的声明，在这种情况下与在线领域。

我需要所有方面（分组，排序和地方），但我不知道如何。

有什么建议吗？的

结果从Galz解释查询：

id select_type   table   type possible_keys key  key_len ref rows Extra 
1 PRIMARY    nomis_houses ref online   online 1  const 8086 Using where 
2 DEPENDENT SUBQUERY nomis_houses index NULL   house_id 4  NULL 9570 Using filesort 

PS。此查询的伎俩，但非常，非常，非常缓慢：

select * from 
(select * from 
    (select * from products order by id desc) tmp_products_1 
group by public_id) tmp_products_2 
where deleted = '0' AND online = '1' 

Answer 1

基于萨钦的答案，你的评论，也许这可以帮助：

select * from products where id in 
(
    select max(id) as id from products 
    where sum(deleted) = 0 
    group by public_id 
) 
and online = 1 

编辑由Pentium10

查询可以改写成

SELECT * 
FROM products p 
     JOIN (SELECT MAX(id) AS id 
      FROM products 
      HAVING SUM(deleted) = 0 
      GROUP BY public_id) d 
     ON d.id = p.id 
WHERE online = 1 

你需要的索引：

• （ID，在线）
• （public_id，删除，ID）

Answer 2

select * from products where public_id in 
(
    select public_id from products  
    group by public_id 
    having sum(deleted) = 0 
) 
and online = 1 

尝试，如果能正常工作。子查询会提供所有未被删除的public_id，然后运行其他联机的筛选器。

Answer 3

嗯，这工作，但我不知道它的效率。基于萨钦的回答

select p.* from products p where p.public_id in 
(
    select p2.public_id from products p2 
    group by p2.public_id 
    having sum(p2.deleted) = 0 
) 
and p.online = 1 
and p.id = (select max(p3.id) from products p3 where p3.public_id = p.public_id);