Question

我有以下查询：

SET @poly = (SELECT `polygon` FROM locations WHERE location_id = 65);

SELECT 
`listings`.`listing_id`
FROM `listings`
STRAIGHT_JOIN `listings_coords` ON `listings_coords`.`listing_id` = `listings`.`listing_id`


WHERE
`listings`.`country_id` = 223
AND `listings`.`active` = 1
AND `listings`.`published_flag` = 1

AND `listings`.`lat` BETWEEN 51.333436956522 AND 51.681263043478
AND `listings`.`lng` BETWEEN -0.40717687879372 AND 0.15165687879372
AND st_within(`listings_coords`.`coords`, @poly)

ORDER BY `created_at` 
LIMIT 10

然而 - 当我为给定的查询执行COUNT（listings。listing_id）时，它会从0.03减慢到4秒！我知道InnoDB在计算时往往比MyISAM慢一点，但这太长了。解释计数：

+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+
| id | select_type | table           | type   | possible_keys                                                                                                                                                    | key                         | key_len | ref                            | rows  | Extra       |
+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+
|  1 | SIMPLE      | listings        | ref    | PRIMARY,listings_country_id_foreign,latlng,listings_country_created_at,listings_compound,listings_lat_lng_country_created_at,listings_country_id_price_qualifier | listings_country_id_foreign | 4       | const                          | 10565 | Using where |
|  1 | SIMPLE      | listings_coords | eq_ref | PRIMARY,coords                                                                                                                                                   | PRIMARY                     | 4       | cp_staging.listings.listing_id |     1 | Using where |
+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+

在查询执行过程中，我是否可以通过任何方式获得该数字？

解释没有计数的查询：

+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+
| id | select_type | table           | type   | possible_keys                                                                                                                                                    | key                         | key_len | ref                            | rows  | Extra       |
+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+
|  1 | SIMPLE      | listings        | ref    | PRIMARY,listings_country_id_foreign,latlng,listings_country_created_at,listings_compound,listings_lat_lng_country_created_at,listings_country_id_price_qualifier | listings_country_created_at | 4       | const                          | 10565 | Using where |
|  1 | SIMPLE      | listings_coords | eq_ref | PRIMARY,coords                                                                                                                                                   | PRIMARY                     | 4       | cp_staging.listings.listing_id |     1 | Using where |
+----+-------------+-----------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------------------------+---------+--------------------------------+-------+-------------+

显示创建表格列表：

CREATE TABLE `listings` (
  `listing_id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `auth_agent_id` int(10) unsigned NOT NULL,
  `country_id` int(10) unsigned NOT NULL,
  `branch_id` int(10) unsigned DEFAULT NULL,
  `property_type` tinyint(4) NOT NULL,
  `listings_status_id` int(10) unsigned NOT NULL,
  `listings_price_qualifier_id` int(10) unsigned NOT NULL,
  `listings_tenure_type_id` int(10) unsigned DEFAULT NULL,
  `listings_class_id` int(10) unsigned DEFAULT NULL,
  `listings_location_type_id` int(10) unsigned DEFAULT NULL,
  `transaction_type` enum('1','2','3') COLLATE utf8_bin NOT NULL,
  `auth_agents_staff_id` int(10) unsigned DEFAULT NULL,
  `published_flag` tinyint(1) NOT NULL,
  `address_1` char(255) COLLATE utf8_bin DEFAULT NULL,
  `address_2` char(255) COLLATE utf8_bin DEFAULT NULL,
  `area` char(150) COLLATE utf8_bin DEFAULT NULL,
  `city` char(255) COLLATE utf8_bin DEFAULT NULL,
  `region` char(255) COLLATE utf8_bin DEFAULT NULL,
  `postcode` char(255) COLLATE utf8_bin DEFAULT NULL,
  `hide_address` tinyint(1) NOT NULL DEFAULT '0',
  `street_view` tinyint(1) NOT NULL DEFAULT '1',
  `lat` decimal(10,6) NOT NULL,
  `lng` decimal(10,6) NOT NULL,
  `listings_coords_id` int(11) NOT NULL,
  `price_sales` decimal(14,2) DEFAULT NULL,
  `price_sales_min` decimal(10,2) DEFAULT NULL,
  `price_lettings` decimal(10,2) DEFAULT NULL,
  `price_lettings_min` decimal(10,2) DEFAULT NULL,
  `price_per_unit_sales` decimal(9,2) DEFAULT NULL,
  `price_per_unit_lettings` decimal(9,2) DEFAULT NULL,
  `min_size` decimal(10,2) DEFAULT NULL,
  `max_size` decimal(10,2) DEFAULT NULL,
  `listings_unit_id` int(10) unsigned NOT NULL,
  `summary` text COLLATE utf8_bin,
  `description` text COLLATE utf8_bin,
  `short_title` varchar(150) COLLATE utf8_bin DEFAULT NULL,
  `title` varchar(200) COLLATE utf8_bin DEFAULT NULL,
  `slug` varchar(200) COLLATE utf8_bin DEFAULT NULL,
  `agent_ref` char(255) COLLATE utf8_bin DEFAULT NULL,
  `active` tinyint(1) NOT NULL,
  `created_at` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `updated_at` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00' ON UPDATE CURRENT_TIMESTAMP,
  `former_id` int(11) DEFAULT NULL,
  `coords` point DEFAULT NULL,
  PRIMARY KEY (`listing_id`),
  UNIQUE KEY `listings_former_id_unique` (`former_id`),
  KEY `listings_price_sales_index` (`price_sales`),
  KEY `listings_price_sales_min_index` (`price_sales_min`),
  KEY `listings_price_lettings_index` (`price_lettings`),
  KEY `listings_price_lettings_min_index` (`price_lettings_min`),
  KEY `listings_country_id_foreign` (`country_id`),
  KEY `listings_agent_id_foreign` (`auth_agent_id`),
  KEY `listings_listings_status_id_foreign` (`listings_status_id`),
  KEY `listings_listings_price_qualifier_id_foreign` (`listings_price_qualifier_id`),
  KEY `listings_listings_unit_id_foreign` (`listings_unit_id`),
  KEY `listings_listings_tenure_type_id_foreign` (`listings_tenure_type_id`),
  KEY `agents_staff_id` (`auth_agents_staff_id`),
  KEY `listings_listings_class_id_foreign_idx` (`listings_class_id`),
  KEY `listings_listings_location_type_id_foreign_idx` (`listings_location_type_id`),
  KEY `listings_listings_branch_id_foreign_idx` (`branch_id`),
  KEY `latlng` (`lat`,`lng`),
  KEY `listings_city` (`city`,`created_at`),
  KEY `listings_country_created_at` (`country_id`,`created_at`),
  KEY `listings_transaction_type` (`transaction_type`),
  KEY `listings_compound` (`country_id`,`active`,`published_flag`,`created_at`),
  KEY `listings_lat_lng_country_created_at` (`lat`,`lng`,`country_id`,`created_at`),
  KEY `listings_country_id_price_qualifier` (`country_id`,`listings_price_qualifier_id`),
  KEY `country_active_published_property_type_lat_lng` (`country_id`,`active`,`published_flag`,`property_type`,`lat`,`lng`),
  KEY `country_active_published_flag_property_type_status_lat_lng` (`country_id`,`active`,`published_flag`,`property_type`,`listings_status_id`,`lat`,`lng`),
  CONSTRAINT `listings_agent_id_foreign` FOREIGN KEY (`auth_agent_id`) REFERENCES `auth_agents` (`auth_agent_id`),
  CONSTRAINT `listings_country_id_foreign` FOREIGN KEY (`country_id`) REFERENCES `countries` (`country_id`),
  CONSTRAINT `listings_ibfk_1` FOREIGN KEY (`auth_agents_staff_id`) REFERENCES `agents_staff` (`agents_staff_id`) ON DELETE SET NULL ON UPDATE SET NULL,
  CONSTRAINT `listings_listings_branch_id_foreign` FOREIGN KEY (`branch_id`) REFERENCES `branches` (`branch_id`),
  CONSTRAINT `listings_listings_class_id_foreign` FOREIGN KEY (`listings_class_id`) REFERENCES `listings_classes` (`listings_class_id`) ON DELETE NO ACTION ON UPDATE NO ACTION,
  CONSTRAINT `listings_listings_location_type_id_foreign` FOREIGN KEY (`listings_location_type_id`) REFERENCES `listings_location_types` (`listings_location_type_id`) ON DELETE NO ACTION ON UPDATE NO ACTION,
  CONSTRAINT `listings_listings_price_qualifier_id_foreign` FOREIGN KEY (`listings_price_qualifier_id`) REFERENCES `listings_price_qualifiers` (`listings_price_qualifier_id`),
  CONSTRAINT `listings_listings_status_id_foreign` FOREIGN KEY (`listings_status_id`) REFERENCES `listings_statuses` (`listings_status_id`),
  CONSTRAINT `listings_listings_tenure_type_id_foreign` FOREIGN KEY (`listings_tenure_type_id`) REFERENCES `listings_tenure_types` (`listings_tenure_type_id`),
  CONSTRAINT `listings_listings_unit_id_foreign` FOREIGN KEY (`listings_unit_id`) REFERENCES `listings_units` (`listings_unit_id`)
) ENGINE=InnoDB AUTO_INCREMENT=34769 DEFAULT CHARSET=utf8 COLLATE=utf8_bin

Answer 1

这条线是罪魁祸首：

LIMIT 10

这就是为什么你的select如此之快。

请删除此行并再次尝试。我认为count版本现在更快。

Answer 2

如果没有SHOW CREATE TABLE，我会做一些猜测......

COUNT是几百甚至几千？这意味着st_within()必须至少执行多次。
lat和lng字段是NULLable？你可以制作它们NOT NULL吗？
latlng索引是（lat，lng），不包括您要过滤的其他三个字段？将它们首先放在索引中可以帮助。（这是唯一真实的答案＆＃34;我正在提供。）
created_at位于listings？
您运行了两次计时以避免缓冲区缓存，并使用SQL_NO_CACHE来避免查询缓存？如果没有，时间可能是假的。

桌子有多大（GB）？ key_buffer_size和innodb_buffer_pool_size的价值是多少？内存多少钱？（如果合适的缓冲区不够大，导致I / O而不是缓存，因此显着减速。）

COUNT比常规选择慢很多

2 个答案: