我在MySQL工作，想要计算购买特定产品的人数。因此，我使用以下查询：
SELECT o.pid AS person, a.article_group_id AS ag, COUNT(o.pid) AS my_count
FROM vda.clustering c
LEFT JOIN vda.orders o ON c.pid=o.pid
LEFT JOIN vda.articles a ON o.aid=a.id
WHERE c.cluster=1 AND c.cluster_round='0'
GROUP BY person, ag
输出正确，但检索此表需要8秒钟。
person ag my_count
1 1 1
7 0 3
13 0 2
13 7 2
...
如果我不计算数字，但只是使用下面的查询检索数据，则只需1秒钟。是否有其他语法我不知道加速计数？
SELECT o.pid AS person, a.article_group_id AS ag
FROM vda.clustering c
LEFT JOIN vda.orders o ON c.pid=o.pid
LEFT JOIN vda.articles a ON o.aid=a.id
WHERE c.cluster=1 AND c.cluster_round=0
person ag
1 1
7 0
7 0
7 0
13 7
13 16
13 36
13 11
13 16
...
编辑：补充信息：
我的表格是使用以下查询创建的。这些文章包含有关产品的信息。订单将人们与他们购买的物品结合在一起。我有单独的表格（此处未显示），其中我将brand_id翻译成适当的品牌。
"CREATE TABLE `persons` ("
" `id` int(8) NOT NULL, "
" `first_name` varchar(64), "
" `last_name` varchar(64), "
" `email` varchar(64) NOT NULL, "
" `gender` enum('M','F','U') NOT NULL, "
" `birthday` date, "
" PRIMARY KEY (`id`), "
" UNIQUE KEY (`email`) "
") AUTO_INCREMENT = 0, ENGINE=InnoDB")
TABLES['articles'] = (
"CREATE TABLE `articles` ("
" `id` int(8) NOT NULL, "
" `article_nr` varchar(20) NOT NULL, "
" `article_group_id` int(8) NOT NULL, "
" `brand_id` int(8) NOT NULL, "
" PRIMARY KEY (`id`), "
" UNIQUE KEY (`article_nr`), "
" FOREIGN KEY (`article_group_id`) REFERENCES `article_groups` (`id`), "
" FOREIGN KEY (`brand_id`) REFERENCES `brands` (`id`) "
") ENGINE=InnoDB")
TABLES['orders'] = (
"CREATE TABLE `orders` ("
" `id` int(11) NOT NULL, "
" `date` date, "
" `pid` int(8) NOT NULL, "
" `aid` int(8) NOT NULL, "
" `price` float NOT NULL, "
" `discount` float NOT NULL, "
" PRIMARY KEY (`id`), "
" FOREIGN KEY (`pid`) REFERENCES `persons` (`id`), "
" FOREIGN KEY (`aid`) REFERENCES `articles` (`id`) "
") ENGINE=InnoDB")
TABLES['clustering'] = (
"CREATE TABLE `clustering` ("
" `pid` int(8) NOT NULL, "
" `cluster` int(8), "
" `cluster_round` varchar(32), "
" PRIMARY KEY (`pid`, `cluster_round`), "
" FOREIGN KEY (`pid`) REFERENCES `persons` (`id`) "
") ENGINE=InnoDB")
具有集群索引的执行计划，cluster_round，如评论中所述：
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE c ref cluster_ci cluster_ci 103 const,const 24624 "Using where; Using index; Using temporary; Using filesort"
1 SIMPLE o ref order_pid order_pid 4 vda.c.pid 1 NULL
1 SIMPLE a eq_ref PRIMARY PRIMARY 4 vda.o.aid 1 NULL