MYSQL加入/索引优化

Question

我有一个查询，试图找到包含一组给定包的所有购物车。 对于每个包，我都会加入相应的cartitem表，因为我只对包含所有给定包的购物车感兴趣。

当我达到超过15个包（连接）时，查询性能迅速下降。

我在相应的外部列上有两个indeces，并且知道mysql只使用其中的一个。当我在2列（cartitem_package_id，cartitem_cart_id）上添加一个索引时，它是有效的，但这是解决这种情况的唯一方法吗？ 我想知道为什么MYSQL突然在这种情况下陷入困境，以及可能是mysql的内部问题，因为我没有看到这个定义和查询有什么更深层次的问题？这可能是查询优化器的问题吗？我可以做些什么（例如添加括号）来支持或强制执行特定的查询吗？或者有任何人在这里使用另一种查询方法？

查询看起来是这样的：

   SELECT cart_id 
      FROM cart 
      INNER JOIN cartitem as c1 ON cart_id=c1.cartitem_cart_id AND c1.cartitem_package_id= 7 
      INNER JOIN cartitem as c2 ON cart_id=c2.cartitem_cart_id AND c2.cartitem_package_id= 8 
      INNER JOIN cartitem as c3 ON cart_id=c3.cartitem_cart_id AND c3.cartitem_package_id= 9 
      INNER JOIN cartitem as c4 ON cart_id=c4.cartitem_cart_id AND c4.cartitem_package_id= 10 
      INNER JOIN cartitem as c5 ON cart_id=c5.cartitem_cart_id AND c5.cartitem_package_id= 11 
      INNER JOIN cartitem as c6 ON cart_id=c6.cartitem_cart_id AND c6.cartitem_package_id= 12 
      INNER JOIN cartitem as c7 ON cart_id=c7.cartitem_cart_id AND c7.cartitem_package_id= 13 
      INNER JOIN cartitem as c8 ON cart_id=c8.cartitem_cart_id AND c8.cartitem_package_id= 14 
      INNER JOIN cartitem as c9 ON cart_id=c9.cartitem_cart_id AND c9.cartitem_package_id= 15 
      INNER JOIN cartitem as c10 ON cart_id=c10.cartitem_cart_id AND c10.cartitem_package_id= 16 
      INNER JOIN cartitem as c11 ON cart_id=c11.cartitem_cart_id AND c11.cartitem_package_id= 17 
      INNER JOIN cartitem as c12 ON cart_id=c12.cartitem_cart_id AND c12.cartitem_package_id= 18 
      INNER JOIN cartitem as c13 ON cart_id=c13.cartitem_cart_id AND c13.cartitem_package_id= 19 
      INNER JOIN cartitem as c14 ON cart_id=c14.cartitem_cart_id AND c14.cartitem_package_id= 20 
      INNER JOIN cartitem as c15 ON cart_id=c15.cartitem_cart_id AND c15.cartitem_package_id= 21 
      INNER JOIN cartitem as c16 ON cart_id=c16.cartitem_cart_id AND c16.cartitem_package_id= 22 
      INNER JOIN cartitem as c17 ON cart_id=c17.cartitem_cart_id AND c17.cartitem_package_id= 23 

输出：

No result. 

考虑下面的示例结构：

CREATE TABLE IF NOT EXISTS `cart` (
    `cart_id` int(10) unsigned NOT NULL AUTO_INCREMENT, 
    `cart_state` smallint(20) DEFAULT NULL, 
    PRIMARY KEY (`cart_id`) 
) ENGINE=MyISAM DEFAULT CHARSET=utf8 AUTO_INCREMENT=80 ; 



INSERT INTO `cart` (`cart_id`, `cart_state`) VALUES 
(1, 0),(2, 5),(3, 0),(4, 0),(5, 0),(6, 0),(7, 0),(8, 0),(9, 0),(10, 0),(11, 0),(12, 0),(13, 0),(14, 5),(15, 5),(16, 10),(17, 0),(18, 10),(19, 40),(20, 10),(21, 5),(22, 0),(23, 10),(24, 10),(25, 0),(26, 10),(27, 5),(28, 5),(29, 0),(30, 5),(31, 0),(32, 0),(33, 0),(34, 0),(35, 0),(36, 0),(37, 0),(38, 0),(39, 0),(40, 0),(41, 0),(42, 0),(43, 0),(44, 0),(45, 40),(46, 0),(47, 0),(48, 1),(49, 0),(50, 5),(51, 0),(52, 0),(53, 5),(54, 5),(55, 0),(56, 0),(57, 10),(58, 0),(59, 0),(60, 5),(61, 0),(62, 0),(63, 10),(64, 0),(65, 5),(66, 5),(67, 10),(68, 10),(69, 0),(70, 0),(71, 10),(72, 0),(73, 10),(74, 0),(75, 10),(76, 0),(77, 10),(78, 0),(79, 10); 


CREATE TABLE IF NOT EXISTS `cartitem` (
    `cartitem_id` int(10) unsigned NOT NULL AUTO_INCREMENT, 
    `cartitem_package_id` int(10) unsigned DEFAULT NULL, 
    `cartitem_cart_id` int(10) unsigned DEFAULT NULL, 
    `cartitem_price` decimal(7,2) NOT NULL DEFAULT '0.00', 
    PRIMARY KEY (`cartitem_id`), 
    KEY `cartitem_package_id` (`cartitem_package_id`), 
    KEY `cartitem_cart_id` (`cartitem_cart_id`) 
) ENGINE=MyISAM DEFAULT CHARSET=utf8 AUTO_INCREMENT=89 ; 


INSERT INTO `cartitem` (`cartitem_id`, `cartitem_package_id`, `cartitem_cart_id`, `cartitem_price`) VALUES 
(1, 4, 2, 200.00),(2, 7, 3, 30.00),(3, 14, 9, 255.00),(4, 14, 9, 255.00),(5, 22, 9, 120.00),(6, 22, 9, 120.00),(7, 13, 13, 300.00),(8, 13, 13, 300.00),(9, 7, 14, 450.00),(10, 13, 14, 250.00),(11, 17, 14, 150.00),(12, 7, 15, 450.00),(13, 13, 15, 250.00),(14, 18, 15, 127.50),(15, 7, 16, 450.00),(16, 17, 16, 150.00),(17, 7, 18, 450.00),(18, 7, 19, 450.00),(19, 17, 19, 150.00),(20, 21, 19, 25.00),(21, 13, 20, 300.00),(22, 7, 21, 550.00),(23, 19, 21, 105.00),(24, 22, 21, 120.00),(25, 17, 22, 150.00),(26, 7, 23, 550.00),(27, 11, 24, 245.00),(31, 7, 26, 450.00),(32, 21, 26, 25.00),(33, 21, 26, 25.00),(34, 22, 26, 120.00),(35, 23, 26, 120.00),(36, 10, 27, 382.50),(37, 22, 27, 120.00),(38, 13, 27, 250.00),(39, 10, 28, 297.50),(43, 7, 29, 550.00),(41, 20, 28, 82.50),(42, 22, 28, 120.00),(44, 7, 30, 550.00),(46, 22, 30, 120.00),(47, 23, 30, 120.00),(48, 21, 18, 25.00),(49, 21, 19, 25.00),(50, 17, 37, 150.00),(51, 17, 37, 150.00),(52, 21, 37, 25.00),(53, 21, 37, 25.00),(54, 4, 45, 1.20),(55, 6, 45, 0.00),(56, 7, 47, 450.00),(57, 4, 50, 200.00),(58, 13, 52, 250.00),(59, 13, 19, 300.00),(60, 9, 19, 0.00),(61, 17, 53, 150.00),(62, 7, 53, 450.00),(63, 22, 18, 120.00),(64, 7, 16, 450.00),(65, 7, 54, 450.00),(66, 7, 57, 450.00),(67, 17, 57, 150.00),(68, 7, 56, 450.00),(69, 17, 59, 150.00),(70, 7, 60, 450.00),(71, 17, 61, 150.00),(72, 17, 63, 150.00),(73, 21, 65, 25.00),(74, 7, 66, 450.00),(75, 7, 67, 450.00),(76, 11, 68, 385.00),(77, 7, 71, 450.00),(78, 11, 73, 385.00),(79, 13, 73, 300.00),(80, 4, 75, 200.00),(82, 7, 73, 30.00),(83, 18, 73, 127.50),(84, 23, 73, 120.00),(85, 7, 73, 30.00),(86, 10, 77, 382.50),(87, 7, 79, 550.00),(88, 17, 79, 150.00); 

给定的查询是一个可能的边缘情况导致在这个例子中没有结果。

   SELECT cart_id 
     FROM cart 
     INNER JOIN cartitem as c1 ON cart_id=c1.cartitem_cart_id AND c1.cartitem_package_id= 7 
     INNER JOIN cartitem as c3 ON cart_id=c3.cartitem_cart_id AND c3.cartitem_package_id= 9 
     INNER JOIN cartitem as c4 ON cart_id=c4.cartitem_cart_id AND c4.cartitem_package_id= 13 
     INNER JOIN cartitem as c5 ON cart_id=c5.cartitem_cart_id AND c5.cartitem_package_id= 17 
     INNER JOIN cartitem as c6 ON cart_id=c6.cartitem_cart_id AND c6.cartitem_package_id= 21 

输出：

cart_id 
------------- 
19 
19 

查询应返回一个包含连接到在这种情况下包（7,9,13,17,21）项的所有车。

Answer 1

我对你的问题的做法是：

SELECT 
    cart_id 
FROM 
    cart 
INNER JOIN 
    cartitem 
ON 
    cart_id = cartitem_cart_id 
WHERE 
    cartitem_package_id IN (7,9,13,17,21)  -- items that got to be in the cart 
GROUP BY 
    cart_id 
HAVING 
    count(distinct cartitem_package_id) = 5 -- number of different packages 
; 

DEMO与您的数据

说明

的原理是用所需的值列表先过滤，在这里你的包。现在计算每个购物车的不同包裹（GROUP BY cart_id）。如果此计数与您的过滤器列表中的值数量相匹配，则每个包必须位于此购物车中。

如果从子查询中获取这些值，则可以用子查询替换IN子句的值列表。

您应该看到，这种方法应该很容易适应类似的需求。